A Corpus of Early Medieval Armenian Inscriptions Author(s): Timothy Greenwood Reviewed work(s): Source: Dumbarton Oaks Papers, Vol.

58 (2004), pp. 27-91 Published by: Dumbarton Oaks, Trustees for Harvard University Stable URL: http://www.jstor.org/stable/3591380 . Accessed: 22/08/2012 21:31
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Dumbarton Oaks, Trustees for Harvard University is collaborating with JSTOR to digitize, preserve and extend access to Dumbarton Oaks Papers.

http://www.jstor.org

A Corpus of Early Medieval Armenian Inscriptions

TIMOTHY GREENWOOD

raditionally the small body of Armenian inscriptions that date from the seventh and eighth centuries has been studied from three scholarly perspectives. These inscriptions have been used extensively by palaeographers, eager to bridge the gap between the traditional date of the creation of the Armenian alphabet by Ma'toc' at the beginning of the fifth century and the earliest dated Gospel from the end of the ninth century.' Architectural historians too have been quick to turn to the epigraphic record, using their chronological precision to trace both the development of Armenian church design and decoration and the history of individual structures.2 Finally, inscriptions in Armenian have been used to attest the presence of Armenian monastic communities in Jerusalem as well as confirming the pattern of Armenian pilgrimage to Sinai and elsewhere.3 The disciplines of Armenian palaeography and Armenian art and architectural history would be far more challenging were it not for the contribution of these inscriptions. Therefore it is perhaps all the more surprising to discover that the historical potential of this body of primary
For use of these inscriptions by palaeographers, see Y.Ta'ean, Aknarkme hnagrutean vray.Ousumnasirut'iwn aruestin (Vienna, 1898); G. Yovsep'ean, Gr 'cut'ean aruestj hin Hayoc'mefjv Kartez hay hnagrut'ean Hayoc'grc'utean 1-11 and A. ev nos. G. 1-8; 1913), (Erevan, 1973), 80-88; pls. Abrahamyan, (Valar'apat, Hayoc'gir gr 'ut'yun M. E. Stone, D. Kouymjian, and A. Lehmann, Album of ArmenianPaleography(Aarhus, 2002), 14, ills. 2 and 3 and 112-15. A marginal note on fol. 463r in the Gospel of Queen Mlk'e, the wife of Gagik Arcruni, king of Vaspurakan, reads: "466 Pages 1 311 Era I Arac' 28 Monday I Areg 6 Thursday"; see Mayr c'uc'ak hayerin ed. B. Sargisean (Venice, 1914), 373-92; M. Janashian, Armenian i Venetik, jeiagrac'matenadaraninMxit'areanc' MiniaturePainting of theMonasticLibrary of San Lazzaro(Venice, 1966), 16-23 and pl. 23; and Stone, Kouymjian, and Lehmann, Album, no. 1. The date of this Gospel remains contentious. Although most authorities favor Armenian Era 311 (25 April 862/24 April 863), this figure has been read variously as 351 (15 April 902/14 April 903) and 361 (12 April 912/11 April 913). Moreover, this marginal note cannot refer to the date of composition of the whole Gospel because the interval between the two specific dates is a mere 38 days. On fol. 137v, a to the monastery of Varag: "I, Mlk'e, handseparate colophon confirms the gift of the Gospel by Queen Mlk'W maiden of Christ and queen of Armenia, in 361 of the era gave this Gospel to this [church of the] Holy Mother of God which, through my own means and expenses, I built for the assistance of me and my king, Gagik, and his children." It need not have been commissioned by her. As Janashian has observed, there can be no doubt that the Gospel was in her possession by 912 at the latest, thereby providing an invaluable terminusante quem for palaeographers. 2J. Strzygowski, Die Baukunst der Armenierund Europa, 2 vols. (Vienna, 1918), 1:29-52; T'. T'oramanyan, Nyut'erhaykakancartarapetut'yan patmut'yan,2 vols. (Erevan, 1942, 1948). For individual use, see N. and J.-M. Thierry, "La cath6drale de Mren et sa decoration," CahArch21 (1971): 43-77. M. E. Stone, "Holy Land Pilgrimage of Armenians before the Arab Conquest," RevBibl 93 (1986): 943 110; idem, TheArmenianInscriptions from the Sinai, Harvard Armenian Texts and Studies 6 (Cambridge, Mass., 1982); idem, "Armenian Inscriptions of the Fifth Century from Nazareth," REArm 22 (1990-91): 315-32.

28

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

material has not been fully appreciated or exploited. Scholarly debate has largely been restricted to reconstructing those dedicatory inscriptions which are fragmentary, with a view to establishing either the date of the foundation or completion of the structure, or the identity of the sponsor.4 Yet on the basis that the inscriptions are authentic, they possess a much broader historical significance. Not only can they operate as secure, independent controls against which to compare the dating mechanisms and social terminology found in contemporary Armenian literary sources; they can also be used to complement these sources because they contain their own reflections of the structure of Armenian society and its evolution under Sasanian, Byzantine, and Islamic hegemony.
THE COMPILATION OF THE CORPUS

The relevant inscriptions and their respective translations are set out in Appendix 1. Complete or partial photographic representations of all but five of these inscriptions are supplied in Appendix 2.5 The corpus is divided into two: those inscriptions originating in Armenia, whether in Armenian, Greek, or Arabic, and those Armenian inscriptions found outside Armenia. As it happens, the latter derive exclusively from Jerusalem. Thus "Armenian" has been defined both geographically and linguistically, while "relevance" has been determined on the twin criteria of date and historical potential. Although such categorization is useful, enabling the body of evidence to be presented in an ordered fashion, it is worth remembering that these divisions are artificial and that it would be unwise to place too much emphasis upon them. They tend to highlight certain common featuresthe location and language of the inscription-at the expense of others, notably content and purpose. Thus several of the mosaic inscriptions from Jerusalem intercede for named individuals in much the same way as several of the Armenian and Greek inscriptions found in Armenia. Conversely, while a majority of the Armenian inscriptions in Armenia commemorate the foundation of a church, a significant minority cannot be confined to a single category, attesting a wide range of contexts and motivations. It is important to stress at the outset that this corpus is not intended to be an exhaustive collection of Armenian inscriptions dating from the early medieval period. Several inscriptions from Armenia have been excluded on the grounds that they are of such a fragmentary nature as to be almost without meaning (Elvard, Ojun, and Djvari).6 Others have been omitted because they comprise little more than a name or group of names (Ptlni,
4 A. A. Manuc(aryan, K'nnutjyun Hayastani IV-XI dareri ?inararakanvkayagreri(Erevan, 1977), 48-106, provides an invaluable summary of the previous scholarship and discusses the suggested readings. 5 The Armenian inscriptions at Hrip'sime (A.2.2), Bagavan (A.5), and Nax'avan (A.8), the Greek inscription at Mastara (A. 15), and the Arabic inscription at Arut (A.20) lack any photographic record. The photographs of the churches of Bagaran (Fig. 3) and Alaman (Fig. 4) include short fragments of the band inscription. The poor condition of several other photographs is self-evident (Figs. 6, 10.4, and 17). 6 Elvard: 7 vols. (Zug, 1980-90), 3: T'oromanyan, Nyut'er,2:127-56; V. L. Parsegian, ArmenianArchitecture, microfiche 008, pls. A9-D2. Ojun: G. Saxkyan, "Ojuni ekelec'u val irjani arjanagrut'yunnere," Lraber HasarakakanGitutyunneri (hereafter Lraber)(1972) 9:90-101; P M. Muradyan, "Ditolutcyunner Ojuni arjanagrut'yunneri vercanut'yan ait'iv,"Lraber(1973) 6:69-81; A. N. Sahinyan, "Ojuni taEari kaiuc'man vamanaki 7: microfiche 159, masin," IFZh (Patmabanasirakan Handis) (1974) 4:123-27; Parsegian, ArmenianArchitecture, REArm 5 (1968): 109-39; armenienne de de M. Mouradian, P C10-D3. "L'inscription l'Fglise Djvari," pls. Djvari: I. Abouladze, "Quelques remarques a propos de l'article de P.Mouradian: L'inscription armenienne de l'6glise de Djvari,"REArm 6 (1969): 373-91; P M. Mouradian, "Encore au sujet de l'inscription armenienne de l'eglise de Djvari,"REArm 6 (1969): 393-411; K. Salia, "Quelques remarques au sujet de deux articles de P. Moura-

TIMOTHY GREENWOOD

29

Sisian, and Ateni) or because an early medieval date is at best unproven (Ani).7 In addition, the corpus does not include the pilgrim graffiti from Sinai and Nazareth, carved by literate but nonspecialist engravers. Although of great interest to palaeographers, this body of material has no significant contribution to make to this study, comprising for the most part a series of Christian names, occasionally combined with a simple invocation. Instead, this corpus is focused upon those inscriptions that clearly date from the early medieval period and possess historical potential, both individually and collectively. Two further preliminary observations need to be made in relation to the composition of the corpus. First, insofar as has been possible, the accepted reading for each inscription has been checked against the photographic record. On the basis of this exercise, several minor modifications have been proposed. Second, each inscription has been reproduced in the Appendix in accordance with conventional epigraphic practice. Distinctions have been drawn between those letters that are certain, those that are probable, and those that are lost or illegible. The latter category is further divided into those that may be restored with confidence and those for which no restoration is proposed. The intention has been to supply a more accurate description of what can and cannot be read. For this reason, the lengthy restorations proposed by I. A. Orbeli, A. A. ManuE'aryan,and S. X. Mnac'akanyan in relation to the identity of the sponsor at Bagavan (A.5) and the date of construction at Mren (A.7) have been omitted.8 These are speculative and have served only to divert attention away from those elements that are secure.
AUTHENTICITY OF THE ARMENIAN INSCRIPTIONS

At the start of the twentieth century, both N. Marr and Orbeli expressed doubts over the authenticity of the earliest Armenian inscriptions in Armenia.9 Their unease was prompted by the presence of numerous vulgarisms and neologisms. Orbeli also had misgivings about the well-developed and well-established letter forms in the inscription at Bagavan (A.5). Their objections seem to have been predicated upon a shared conviction that the Armenian language was originally uniform in nature. Whereas the written form of the Armenian language may have possessed such a quality in the immediate aftermath
dian: 1. Sur la critique de texte de la troisieme lettre du Catholicos Abraham; 2. Linscription arminienne de l'glise de Djvari,"BK 26 (1969): 145-60. SkP7 Ptini: Parsegian, Armenian Architecture, 1: microfiche 039, pl. E3: rUlO [b]-tjI iruILSYIh1"6&U "Manuel lord (tir) of Amatunik'." Ateni: Z. N. Alek'sidze, Atenis Sionis somxuricarcarebi(Tiflis, 1978), 119-23 and pl. 1: OU ' UPr'WPM fA~l4UI UfAL Pf b[6fMU]--"I T'odosak, builder of this holy church." Sisian: Manuc'aryan, 80-82: 0 95 S SkP 104U0 ULhh Hayastani, UlibbL UlllU 08 StP Lord of Siwnik'/Lord 'llUflUS l I K'nnut'yun lh 0 6b UU'I h9-"Kohazat Yovsep' bishop of fiaf)lu uPlU UPhr AUlCn~cbll/ 6.U sacristan." Ani: YovsEp'ean, K'artiz, no. 4 and pl. 4; I. A. Orbeli, Siwnik'/T'odoros monk of Sion/Yovhanes Izbrannye trudy (Erevan, 1963), 413-19 and pl. LVII: OU rULP'UQ S Chtlb1' .?'R04 hb1U Ph UPUIrf. .. Absolm built this in church Era of the Armenians .. ."This in41 .]-"I vardapet AiTNUIhUllhAbUlk' [10 as referring to Era 71, but it seems more probable that 4.1 are in fact the first scription has been interpreted two letters of Hayoc'. trudy,371-87; S. X. Mnac'akanyan, "Bagavani taEari Sinararakan arjana8 Bagavan (A.5): Orbeli, Izbrannye vercanman harc'i IFZh (1964) 2:213-26. Mren (A.7): A. A. ManucEaryan,"Mreni ew T'alini grut yan iurje," ta'arneri linararkan arjanagrut'yunnere," IFZh (1966) 1:247-52; S. X. Mnac'akanyan, "Erb e karuc'vel Mreni tacare," IFZh (1969) 3:149-64. 9 C'uc'ak For Marr'sobjections, see K. Kostaneanc', VimakanTaregir. (St. Peterszotovacoyarjanagrut eanc'Hayoc' 371-404. xviii-xix; Orbeli, 1913), burg, Izbrannyetrudy,

30

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

of its final elaboration, it seems inherently unlikely that contemporary speech was so characterized, nor that such uniformity would have persisted even in the written form for any length of time. Instead of dismissing these inscriptions on the grounds of linguistic nonconformity, it seems preferable to accept them as good evidence for contemporary variation in both syntax and spelling.'0 However, their arguments do raise a key issue, namely, the authenticity of these inscriptions. It is not enough simply to dismiss the possibility that anyone might wish to commission a false inscription, nor to query the means and opportunity such an operation would require. In order to gauge whether or not an inscription is genuine, two diagnostic tests have been undertaken. The location and layout of each inscription have been established and the results compared with one another and with other Near Eastern epigraphic traditions. Six of the Armenian inscriptions from Armenia can be accepted as original, and hence authentic, solely on the basis of their position and layout. The second test has been palaeographic, requiring the forms of the letters to be consistent with an early date. In his study of Armenian church architecture, J. Strzygowski surveyed those inscriptions he considered to be relevant to his project and divided them into three categories: and "Mehrzeilige Flichan-[oder Felder-]inschriften "Giirtelinschriften,""Ersatzinschriften," mit urspriinglicher Schrift," that is, band inscriptions, substitute inscriptions, and multiline flat or panel inscriptions in original script."lAlthough his controversial theories on the relationship between Armenian and European church architecture have long since been challenged, his preliminary study of the epigraphic evidence remains an important starting point, not least for the inclusion of several photographs that illustrate the position and layout of certain inscriptions on structures now completely destroyed.12 Under his first heading, Strzygowski discussed two band inscriptions in detail, namely, those at Bagaran (A.3) and Alaman (A.4), and noted the existence of two others, at Elvard and Nax'avan (A.8).13 In fact, six inscriptions should be included under this heading: the four identified by Strzygowski, together with the inscription at Bagavan (A.5) and the fragmentary inscription at Ojun. Four of these inscriptions are included in this corpus: those at Bagaran (A.3), Alaman (A.4), Bagavan (A.5), and NaxEavan (A.8). Those at Elvard and Ojun have been omitted from the corpus because they are extremely fragmentary and so fail the "historical potential" test. Nevertheless, they have an important contribution to make to this study of location and layout, not least because the other four inscriptions no longer exist. The following descriptions have been pieced together from surviving records and photographs. According to Orbeli, the inscription at the church of Bagaran (A.3) encircled the exterior of the whole building, beginning at the northern face of the western apse and running across the southern, eastern, and northern apses respectively.'" Figure 3 confirms that the inscription was carved as a single line of text onto the topmost row of prepared masonry
rather than hat'oikosut'eanand episkaposi rather than episkoposi at Tekor (A.1); and See, e.g., kat'aiikosut'ean at rather than marcpanut'ean marzpanut'ean Bagaran (A.3). 11Strzygowski, Die Baukunst, 1:32. 12 Strzygowski's theories of artistic development are scrutinized in C. Maranci, MedievalArmenianArchitecture:Constructions of Race and Nation, Hebrew University Armenian Studies 2 (Leuven, 2001). 13 Strzygowski, Die Baukunst, 1:32-38. 14 Orbeli, Izbrannye trudy,390.
10

TIMOTHY GREENWOOD

31

blocks, located immediately below the level of the roof. Although Orbeli's reconstruction indicates that the inscription ran as a single line across the majority of the blocks, it also reveals that several blocks carried double lines of text.'5 Two lines of text were consistently applied around the northern apse which carried the last part of the inscription. This switch from a single band to double lines suggests to me that the carver ran out of space. In order to complete the inscription, he was forced to revise the relationship between text and surface and alter the layout. However, this only invites the question: why did he not plan the inscription to fit the area available? The most attractive proposition is that the inscription at Bagaran is composite, comprising two distinct parts. The first ran around the western, southern, and eastern apses. The second was limited to the surface of the northern apse, where the switch to double lines of text occurred. This proposition is supported by significant differences in content between the two parts. The first part of the inscription is chronologically precise. It identifies the original sponsor as But AIuelean as well as revealing how his wife, Annay, took over the responsibility for the completion of the construction after his murder. The second part is undated. Its meaning is by no means clear, but it appears to be limited to intercession on behalf of Annay and her young children and on behalf of Su'an, who may be Annay's sister. Thus it seems to me that the two parts are directly linked to each other through the person of Annay. On this basis, I would conjecture that the wording of the inscription was expanded at the time of completion or shortly afterwards at the instigation of Annay herself, prompting the use of double lines on the northern apse to accommodate the additional intercessions. Intriguingly, Annay seems to be inviting God's protection over herself and her children in this world rather than interceding on behalf of their souls in the next. Could it be that her husband's murder prompted Annay to make this addition to the inscription, as she struggled to defend the interests of her offspring? In contrast to the above, the band inscription on the church at Alaman (A.4) extended across only the southern and eastern elevations, ending on the northeastern wall of the eastern apse.16 Since the wording of the text comes to a natural conclusion at that point, it seems that the northern and western exterior walls were left blank intentionally. According to Orbeli, the inscription was carved as a single line of text across a single row of blocks without recourse to twin lines at any point." Strzygowski's photographic record has again proved invaluable in confirming its location. Whereas the foundation inscription at Bagaran (A.3) was located on the highest row of prepared blocks below roof level, Figure 4 reveals that the inscription was located on the fourth row below this level, even though this caused the linear continuity of the inscription to be interrupted by two small windows. It is worth noting that neither of these photographs was taken with the intention of recording the inscription. That both do so confirms that the inscriptions were legible from some distance away and from ground level and implies that they were intended to be so. For reasons unknown, the inscription at Bagavan (A.5) was not included by Strzygowski in his category of band inscriptions. However, Atilan records that the inscription ran in a single line around the southern, eastern, and northern elevations of the church.'8 This is partially confirmed by Orbeli. His painstaking reconstruction indicates that the
15 Ibid., 391.
16 17 18

Ibid., 394-95; Strzygowski, Die Baukunst, 1:36. Orbeli, Izbrannye trudy,395, 397. L. Atidan, AyraratbnasxarhHayastaneayc' (Venice, 1890), 529.

32

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

inscription ran along the southern fagade and then across the external faces of the three apses along the eastern elevation. However, he does not indicate that it extended onto the northern wall as well.'9 Unfortunately the photographs of this large church supplied by Orbeli are of very poor quality and it is not possible to make out the height at which the inscription ran.20 The inscription from the church at Nax'avan (A.8) was already incomplete when it was first recorded. According to Atisan, a single line of text ran as a band across the exterior surface of the north wall, although Manui'aryan also refers to characters on the surface of the western elevation as well.21 Both the beginning and the conclusion of the inscription are missing, implying that at least two and possibly all three of the other elevations were similarly girdled. Strzygowski's photograph is of very poor quality, but it may be possible to make out the faint imprint of the inscription two-thirds of the way up the wall, on the upper part of the fourth row of blocks beneath the roofline, pierced once again by windows. If this were correct, its location would be similar to the inscription at Alaman (A.4). This study of the location and layout of band inscriptions must also take into account the evidence of the inscriptions at Elvard and Ojun. Although the inscription stretching along the southern wall of the church at Elvard is fragmentary and difficult to read, it too comprises a single band of text. It was carved at a height just below the lintels of the three door frames, extending across the portal capitals. It is therefore at a slightly lower level than those described above.22 Tc. Tcoramanyan dated the construction of this church to the later sixth century because he believed that the inscription referred specifically to the katcolikos Movsis II (574-604).23 Later commentators have challenged the accuracy of his reading. Unfortunately the issue is destined to remain unresolved because the relevant part of the inscription has suffered considerable weathering and is no longer legible. A. Sahinyan also modified Tcoramanyan's argument on architectural grounds.24 In spite of these differences, there seems to be little doubt that the band inscription is either fifth- or sixth-century in date. The church at Ojun also preserves several fragments of a band inscription. These are located on its eastern elevation, four rows up from ground level. Unlike the other inscriptions, it is not possible to establish whether or not this inscription originally ran in a continuous band along this elevation. Although at least two blocks carry very well preserved letters, several of the intervening blocks do not bear the trace, however weathered, of any carving. It is clear, however, that the block bearing the name "T'eodoro[s]" has been recut, enabling a block carved with a xa'kar (carved medieval cross) to be fitted beneath it.25Thus although those blocks carrying letters appear to derive from an early medieval band inscription, it is not certain that they are in their original location. The proposition that these blocks were moved from their original location in the course of a rebuilding would explain the low position of this inscription as well as its unique location, confined exclusively to the eastern elevation.
19 20 21 22 23 24 25

Orbeli, Izbrannyetrudy,372-74. Ibid., pls. LXV-LXVII. L. Atisian, Siraktetagrut'iwn patkerac'oyc'(Venice, 1881), 136; Manuicaryan, K'nnut'yun Hayastani, 90. 1: Die 35. Baukunst, fig. Strzygowski, T'oramanyan, Nyut er, 2:127-56. A. Sahinyan, Kahsati (Erevan, 1955), 101-14. bazilikayi cartarapetut'yun? Parsegian, ArmenianArchitecture,1: microfiche 021, pls. C9-D5.

TIMOTHY GREENWOOD

33

In spite of these reservations, this group of band inscriptions possesses several common features. Each comprises a single line of text which stretched counterclockwise around the exterior walls of the church. The surviving photographs indicate that the inscriptions were legible from ground level, implying that this was by design. Those that can be deciphered all commemorate the foundation of that church by a named individual in a specified year. It is important to stress that these churches were located in different districts of historic Armenia, that they were constructed at different times during the early medieval period, and that their inscriptions commemorate unrelated sponsors. Therefore their common characteristics cannot be explained as a regional style, as the product of a particular time or a particular sponsor. Collectively they attest a long-standing tradition in Armenia of carving band inscriptions at a high level on external walls. By virtue of their location and layout-and thus independently of their content-these inscriptions should be accepted as original. Frustratingly, efforts to trace the origin of this tradition have proved largely fruitless. Exterior band inscriptions are not found on churches across the Near East in late antiquity. The vast majority of foundation and other inscriptions are confined to door lintels or mosaic floors. Of course, one should remember that there are few late antique structures which survive to a sufficient height and that arguments from silence are at best inconclusive and at worst specious. Nevertheless, the dearth of corresponding inscriptions outside Armenia is striking. Two sixth-century band inscriptions found in Constantinople appear to constitute the nearest contemporary equivalents. The excavation of the church of St. Polyeuktos (ca. 524-527) revealed that the first half of the dedicatory epigram (lines 1-41) was carved as a single line of text at the height of the entabulature on both sides of the nave.26 The second half of the epigram (lines 42-76) was carved in the narthex on five separate tablets. The church of Sts. Sergius and Bacchus (ca. 527-534) likewise contains a single-line inscription encircling the interior of the church at the level of the entabulature.27 In terms of location and content, however, these two inscriptions are far removed from the Armenian tradition. Moving ahead to the end of the seventh century, both the outer and inner faces of the octagonal arcade within the Dome of the Rock in Jerusalem carry original single-line Arabic inscriptions in mosaic. That on the outer face refers to the year A.H. 72 (691/692), although scholarly opinion remains divided on the question of whether this is the year in which construction began or was completed.28 These inscriptions therefore reveal something of the interaction between viewer and layout and the context in which the inscription should be interpreted. The text can be read only by walking around the outer and inner faces. In this instance, the layout of the inscription dictated ceremonial movement within the building. Again the Armenian tradition can be distinguished on the basis of location and content. In this instance, however, it is worth noting
R. M. Harrison, Excavationsat Sarafhane in Istanbul, 2 vols. (Princeton, 1986), 1:3-8, 117-19, 407-11 and 91, and 93-105. 87-88, pls. 27 C. Mango, "The Church of St. Sergius and Bacchus at Constantinople,"JOB 21 (1972): 190. 28 S. S. Blair, "What Is the Date of the Dome of the Rock?" in Bayt al-Maqdis, SAbd al-Malik's Jerusalem,Oxford Studies in Islamic Art 9, part 1, ed. J. Raby and J. Johns (Oxford, 1992), 59-87. The inscription on the outer face concludes with the statement: "The servant of God built this dome 'Ab[d al-Malik commander] of the faithful in the year 72, may God accept it from him and be pleased with him. Amen. The Lord of the Worlds. Praise belongs to God."
26

34

A CORPUS OF EARLYMEDIEVAL ARMENIAN INSCRIPTIONS

that the exterior of the Dome of the Rock originally carried mosaics as well.29 The remains of these were replaced with tiles by Sulayman the Magnificent between 1545 and 1552. Sheila Blair has suggested that one of their purposes may have been to direct pilgrims to enter through the south door, where they would be confronted with the opening section of the inscription on the outer ambulatory.30Such a message, however, need not have been communicated via a band inscription or indeed in verbal form, and the reconstruction of the original exterior proposed by H. R. Allen does not provide for an inscription.31 The earliest example of a non-Armenian external band inscription appears to be the Georgian inscription carved in the third quarter of the eighth century on the exterior of the church of Sion at Samivilde (Samiwlde), southwest of Tblisi.32Although the church is located in a district of historic Iberia rather than Armenia, it is less than fifty miles from Ojun, whose Armenian band inscription was discussed previously. A direct link between the Armenian tradition and this Georgian inscription seems very likely, even if the nature and circumstances of that connection remain obscure. By contrast, the earliest Byzantine external band inscription is much later, dating from the second half of the ninth century. The external face of the central apse of the church of the Virgin at Skripou contains a series of eight panels in a single continuous band, inscribed with the prayer of the donor, Leo basilikos and epi ton oikeiakon,to the Virgin.33This foundation is dated by an protospatharios exterior panel inscription on the north wall to the year 6382 since the creation of the world, which equates to the year 873/874.34 A second, separate inscription, commemorating the dedication of a church to the Mother of God by Constantine Lips in Constantinople in June 907, fits the same pattern. It was located on a beveled marble cornice which girdled the exterior of the three central apses of the northern church of this monastic complex.35 Thus both these Greek band inscriptions were confined to the external surface of the apse. When the gulf of some two centuries is also considered, a direct association between the Armenian and Byzantine traditions becomes less rather than more probable. In late antiquity, the external band inscription stretching across two or more elevations constituted a uniquely Armenian tradition. The inspiration for the external band inscription will always be a matter of speculation. However, several observations may be advanced. First, this epigraphic tradition is consistent with the unmistakable Armenian preference for exterior carved decoration,
29 Blair, "Date of the Dome of the Rock," 70; H. R. Allen, "Observations on the Original Appearance of the Dome of the Rock," in Bayt al-Maqdis.Jerusalemand Early Islam, Oxford Studies in Islamic Art 9, part 2, ed. J. Johns (Oxford, 1999), 197, 208-9. 30 Blair, "Date of the Dome of the Rock," 76. 31 Allen, "Observations," 208-9 and color pl. at 202-3. 32 N. Chubinashvili, SamshvildskiiSion (Tbilisi, 1969), 28 and n. 35. Chubinashvili reproduces the reconstruction proposed by L. Muskhelishvili, "Nadpisi Samshvildskogo Siona i vopros vremeni ego postroeniia," IzvestiiaInstituta lazyka Istorii i MaterialnoiKulturyim Akad. N. Ia. Marra 13 (1942): 92. According to Muskhelishvili, the total length of the inscription around the fagade originally exceeded 26 m. I am very grateful to Professor J.-P Mah6 for drawing my attention to this inscription. Professor R. W. Thomson has kindly supplied an English translation of this reconstruction; see note 106 below. 33 N. Oikonomides, "Pour une nouvelle lecture des inscriptions de Skripou," TM 12 (1994): 479-93 and pls. I-iv; A. Papalexandrou, "Text in Context: Eloquent Monuments and the Byzantine Beholder," Word& Image 17.3 (2001): 266-67 and figs. 5 and 6. 34 Oikonomides, "Skripou," 482-83 and pl. III. 35 T. Macridy, A. H. S. Megaw, C. Mango, and E. J. W. Hawkins, "The Monastery of Lips (Fenari Isa Camii) at Istanbul," DOP 18 (1964): 254-55, 300-301, and fig. 1.

TIMOTHY GREENWOOD

35

particularly around the frames of both windows and doors.36 Indeed, one could argue that the horizontal band of text gives a certain unity to the exterior of the building, linking the separate decorated surfaces. This is most obvious on those structures whose windows or door frames break up the continuity of the text. Second, since these are all foundation inscriptions, there can be no doubt that their location and layout were planned. Their position was not dictated by existing inscriptions or decoration. Instead it reflects a deliberate choice. This leads to the third observation. At first sight, the single band of text is not an obvious mechanism for a public declaration of ownership and responsibility. It requires the viewer to walk around the structure. Therefore, as at the Dome of the Rock, the location and layout of the inscription tell us something about the demands upon, and the response expected from, the literate viewer or accompanying listener. This prompts the question: in what circumstances would anyone walk around the exterior of the church reading the whole inscription, and so, by implication, praying to God for intercession?37 In my view, it is very significant that two of the four band inscriptions, those at Bagaran (A.3) and Bagavan (A.5), as well as two other foundation inscriptions, at Mastara (A. 10.4) and Aru' (A. 11), all specify the day and the month on which the church was completed.38 This pattern suggests that such inscriptions may have played some role in an annual service of commemoration for the founders of the church during which a circumambulation of the building took place. Such a service would have fulfilled the ongoing mediatory role of the building on behalf of the souls of its founders. Unfortunately this attractive solution is undermined by the inscription at Alaman (A.4) which does not specify any particular month or day. Moreover, it does not explain the height of the band inscriptions. Was this intended to have a purely practical benefit, protecting the inscription from defacement in the future? The fragmentary nature of the relatively low inscriptions at Elvard and Ojun supports this proposition. Was it intended to elevate the inscription above the world, and by implication closer to God, who would be reminded of the founders' concern for their souls when contemplating the community of the faithful assembled within? Or was it a tradition which derived from similar practices in relation to smaller precious objects and
The biblical inspiration for this decoration has been noted: see P Donabidian, "Themes bibliques et sculpture pr6-arabe" REArm 22 (1990-91): 253-314, esp. 275-82; and E B. Flood, The GreatMosqueof Damascus:Studieson theMakings of an UmayyadVisualCulture(Leiden, 2001), 57-113, esp. 71 and 79, where the biblical precedents behind Armenian vine and pomegranate exterior carved decoration are traced. See, for example, Isa. 5:7 which explicitly associates the children of Israel, the people of God, with the vine. 37 Papalexandrou, "Text in Context," 274-83. 38 Bagaran (A.3): 20 TrE; Bagavan (A.5): 30 Hrotic'; Mastara (A.10.4): 14 Arac'; Aru' (A.11): 15 Mareri. By way of comparison, the History of Yovhannes Drasxanakertc'i contains four similar dates, specifying a particular day and month: see Yovhannes Drasxanakertc'i Uohn Catholicos], Yovhannukat'otikosi Drasxanakertec'woy Patmut'iwnHayoc',ed. M. Emin (Moscow, 1853; repr. Tiflis, 1912; Delmar, N.Y., 1980), 125, lines 9-10 (25 Mehekan), 254.10-11 (25 Mareri), 350.27-351.1 (10 Ahik), and 353.12 (17 Ahik); trans. and commentary by K. H. Maksoudian, Yovhannes Drasxanakertci. HistoryofArmenia, Occasional Papers and Proceedings 3 (Atlanta, 1987), 25.70, 51.42, 66.62, and 66.73. These precise dates are all found at the conclusions of accounts detailing the sufferings of faithful Armenian martyrs. It seems to me very likely that such passages were included in, and derived from, the annual liturgical cycle. This would justify the inclusion of a specific date (for the purposes of commemoration) as well as their intriguing combination of vivid, self-contained narratives with lengthy prayers. These narratives were preserved and transmitted through the liturgy as Armenian examples of Christian fortitude in the face of oppression. Specific dates appear in other medieval Armenian historical texts in exactly the same position and context, implying that the liturgical cycle was used as a source of historical information by contemporary writers.
36

36

A CORPUS OF EARLYMEDIEVAL ARMENIAN INSCRIPTIONS

manufactures? Although the reason for their elevated position remains unknown, it seems very likely that the band inscription had a role in recording the exact date on which the church was completed and hence the day on which its founders were to be commemorated each year. Such celebrations would have been particular to the individual church and need not have been reflected in the annual liturgical cycle of the Armenian Church. One final observation should be made on the location of the band inscriptions. There appears to be a broad coincidence between those surfaces carrying the band inscriptions and those that are pierced by doorways. At Elvard it is the southern elevation, possessing no fewer than three separate entrances, that carries the inscription. At Bagaran (A.3) the first part of the inscription stretched around the western, southern, and eastern elevations in that order; the entrances into this church were to be found along the western and southern walls. The inscription at Alaman (A.4) covered the southern and part of the eastern elevation; the southern entrance seems to have been the principal entrance, judging by its size and decoration in comparison to the western doorway. At Bagavan (A.5) the inscription extended across the southern and eastern sides and may have extended onto the northern elevation as well; this church had five entrances, two of which were located on its southern and northern sides. On the basis of these examples, the southern elevation seems to be favored in Armenian church design, both in terms of access into the church and in the location of the band inscription. Although conjectural, this preference may reflect the presence of adjacent structures to the south of the church, structures that currently await discovery or further excavation. At the present time, little is known of the architectural context in which many of the surviving Armenian churches were originally situated, but this might explain this preference. It is significant that the partial excavation of the area around the church at Aru' has revealed a palace located immediately south of the church, aligned to facilitate access between the two.39 Before advancing to consider Strzygowski's second category, "Ersatzinschriften" or "substitute inscriptions," we should note that there is another group of inscriptions which should be accepted as original by virtue of their position. It comprises those foundation inscriptions located on the western facade of each church immediately below the level of the roof, carved as three or four lines of text onto a single row of blocks. Two inscriptions qualify, those at Mren (A.7) and T'alin (A. 12). The Mren inscription consists of three lines of text, whereas that at T'alin is spread across four lines. It is very unlikely that anyone would choose to carve unrelated inscriptions in this similar, difficult location onto different buildings other than at the time of original construction. Indeed, the damage susconxa-kars tained to the inscription at Mren through the subsequent carving of medieval firms the priority and antiquity of the inscription. Once again, both inscriptions appear to be directly associated with the main entrance of each church. The lintel of the western doorway of the church at Mren is decorated with an elaborate relief, depicting the Armenian donors of the church flanking Christ, St. Peter, and St. Paul.40 This western orientation distinguishes these two inscriptions from the band inscriptions whose southerly preference was noted above. It also links them to the Byzantine epigraphic tradition, which favored the western elevation. The western orientation of the main entrance and
3 A. Khatchatrian, L'architecture arminienne du IVeau VIesizcle, Bibliothbque des Cahiers Archdologiques 7 (Paris, 1971), 63-64 and fig. 75. 40 N. and J.-M. Thierry, "La cath6drale de Mren," 57-69 and figs. 13-19.

TIMOTHY GREENWOOD

37

foundation inscription at both these churches may therefore reflect the political-and perhaps the doctrinal-affiliation of their founders. Both inscriptions attest an impressive array of prestigious Byzantine titles bestowed upon the respective founders. Moreover, the relief carved onto the lintel of the northern doorway of the church at Mren depicts the return of the True Cross to Jerusalem by Emperor Heraclius on 21 March 630, an overtly imperial theme without any obvious Armenian connotation.41 When considered together, these elements suggest that the church at Mren was intended to reflect its sponsor's support for Heraclius. However, in the context of Armenian dedicatory inscriptions, it should be noted that this western orientation is atypical. Indeed, traditional Armenian practice persisted into the eleventh century. The original foundation inscription of the cathedral at Ani commemorating the benefaction by Queen Katranide and the completion of the building in 1001 is located on its southern fagade.42 Intriguingly the later public inscription commissioned by the Byzantine magistrosand katepan of the East, Bagrat Vxkac'i, in the reign of Constantine X Doukas (1059-67) appears on the western elevation.4" This clear distinction between the Armenian and Byzantine epigraphic traditions in relation to the preferred orientation of the inscription endured into the eleventh century. In accepting the foundation inscription at Mren (A.7) as original and placing it in a separate group, together with the inscription at T'alin (A.12), Strzygowski's second category, "substitute inscriptions," is reduced to two examples, those at Tekor (A. 1) and Mastara (A. 10.1-4). The former was located in the architrave above the tympanum of the western entrance of the church. However, as Strzygowski and others have observed, the inscription is not only cramped but also can only be understood when read in reverse, from the last line of text upward.44A. Khatchatrian concluded: "La seule explication de cette inversion de l'inscription et de son insertion maladroite sur l'architrave est que, dans l'ftat oii elle s'etait conservee, elle prdsentait comme les [autres] inscriptions du tympan, une copie de l'inscription originale, execut e en dernier lieu."45Furthermore, a close reading of the inscription confirms that it is in fact composite, being made up of two separate elements. The first refers to Sahak Kamsarakan building a vkayaranor martyrium of St. Sahak, asserting a mediatory function for him and his wife and family; the second describes how this tetis or site was founded by Yohan the kat'olikos of Armenia in conjunction with three other named individuals. According to Tcoramanyan, these two parts cannot be combined within a single original inscription and should therefore be interpreted as separate inscriptions deriving from separate structures.46 The inscriptions examined hitherto are all singular in the sense that each occupies a single unbroken sequence of blocks. The Armenian inscriptions found on the church at Mastara deviate from this pattern. No fewer than four separate inscriptions collectively attest the dedication of the church (A.10.1-4). Although they consistently identify one
41 Ibid., 69-76 and figs. 28-33. See also N. Thierry, "Heraclius et la vraie croix en Armenie," in FromByzantium to Iran: ArmenianStudiesin Honour of Nina G. Garsoian,ed. J.-P. Mahe and R. W. Thomson (Atlanta, 1997), 165-86. 42 Divan hay vimagrut'yun (hereafter DVH), ed. H. A. Orbeli, S. Barxudaryan, and S. A. Avagyan, 7 vols. (Erevan, 1960-82), 1: no. 101 and pl. xiii. 43 DHV 1: no. 106 and pl. xII. 44 Strzygowski, Die Baukunst, 1:39-41. arminienne, 50. 4 Khatchatrian, L'architecture 46 Tcbramanyan, Nyut'er,1:218-20.

38

A CORPUS OF EARLYMEDIEVAL ARMENIAN INSCRIPTIONS

Grigoras as the founder of the church, they are found at different places on the exterior. Following the sequence used by Manuc'aryan, the first (A.10.1) is carved onto a single block located immediately above the window over the entrance into the southern apse of the church, although the inscription also continues around the semicircular frame of the window.47The second (A.10.2) is found in the upper register of the southern section of the western elevation, again above the level of the window. Seven lines of text cover three rows of blocks. Once again, it appears that the engraver encountered difficulty fitting the end of the inscription into the available space. He therefore carved two lines of text onto the southern face of the western apse, perpendicular to the original surface, at the level of the bottom row of carved blocks. In other words, the carver was forced to move across to the surface adjacent and to the left of the original surface in order to complete the inscription. The third inscription (A. 10.3) is located above the window and the doorway of the western apse. It is carved onto two blocks, one on top of the other. The six lines of text on the upper block have been squeezed into the space on either side of a carved cross, suggesting that the cross predates the inscription. The second block is much longer and carries five lines of text, but these again appear cramped. Finally, the fourth inscription (A. 10.4), supplying the month and day of the consecration of the church, is positioned on the western section of the southern elevation. It comprises four lines of text carved onto a single block, and one further line of text, carved onto the adjacent block, level with the bottom line of text on the previous block. Unlike the three other inscriptions, this is clearly incomplete because it lacks the name of the bishop who was commemorated on the same day as the consecration of the church. How should we interpret the location of these inscriptions on the church at Mastara? They are concentrated around the two entrances to the church and the external vertical surfaces between them. Furthermore, they are located high up on the exterior, above the level of the windows. In these respects, one can detect clear correspondences with the groups of inscriptions described above. Their layout is more difficult to explain, appearing cramped and hard to read. These features are typical of substitute inscriptions. Yet it is hard to envisage the circumstances in which they were all recarved with such diligence onto separate inaccessible surfaces. Although the three principal Mastara inscriptions appear poorly planned and often compressed, I am not convinced that this is sufficient reason to dismiss them as substitute. Their originality must remain an open question. The authenticity of the wording is less contentious. Although the use of four separate and different inscriptions to commemorate the foundation of the church has not been encountered previously, this practice should be compared with the two separate and different inscriptions recording the role of Komitas in the construction of the church of St. Hip'sime in EJmiacin (A.2.1-2). The first of these is located on the exterior surface of the western elevation and refers to Komitas as a sacristan of St. Hiip'sime who was summoned to the throne of St. Grigor; it therefore attests his role in the construction of the church. The second is positioned on the interior surface of the eastern elevation and refers to Komitas as kat'olikos of Armenia; it also invokes Christ, asking him to remember Komitas.48Thus the two inscriptions have different temporal dimensions and fulfill different functions. The first is pragmatic, recording the past career of Komitas, his appointment, and his personal responsibility for the construction; the second presents Komitas as the
47 48

ManuC(aryan,K'nnutyun Hayastani, 53-54.

Ibid., 77-80.

TIMOTHY GREENWOOD

39

current kat'olikos and implores Christ to recognize his labors. In much the same way, each inscription at Mastara has its own unique purpose. The first (A.10.1) intercedes for Grigoras as well as dating the construction of the church through the association with Bishop T'eodoros. The second (A.10.2) expresses Grigoras's thanks to God for granting him the opportunity to build the church, as well as his subsequent appointment as bishop of Apahunikc.49It also defines the function of the church in four different ways: as a refuge for Mastara itself, as a place of prayer for the faithful, a place of expiation for sinners, and a memorial for Grigoras and his unspecified relatives. The third (A. 10.3) recalls the relationship between Christ and his church, portrayed as the bride of Christ.50 It also refers obliquely to the dedication of the church to St. Yovhannes, the apostle, prophet, and martyr, who is the "companion of the bridegroom," Christ himself. Finally, the fourth inscription (A. 10.4) records the month and day of consecration, thereby setting the date for an annual service of commemoration. Collectively, therefore, these inscriptions reveal a wealth of information about the foundation of the church at Mastara. They attest a tradition of multiple dedicatory inscriptions which need to be read together in order to obtain the fullest meaning. The two inscriptions at the church of St. Hiip'sim4 belong to the same tradition. The third and final category proposed by Strzygowski, "Mehrzeilige Flachan-[oder Felder-]inschriften," has been reduced by the earlier separation of the inscriptions at Mren (A.7) and T'alin (A. 12) into their own category. The solitary panel inscription of the early medieval period is found inserted into the eastern wall of the church at AruE (A. 11). Although doubts were expressed by both Orbeli and Strzygowski over the originality of this inscription, on the grounds that it refers to the seemingly impossible "29th year of Constans," this is not sufficient reason to dismiss it.51 In much the same way that silver coins bearing the name and representation of Yazkert III persisted in the former Sasanian Empire long after his death, so it is conceivable that a dating mechanism based on the regnal years of Constans II continued to be employed after his murder in distant Sicily, especially if there was no alternative annual cycle in use in Armenia at that time.52 The unusual dimensions of the single panel, different from those of the surrounding revetment leaves, support the contention that it is original. The remaining Armenian inscriptions from Armenia are scattered across different surfaces in various ways and do not conform to any of the above patterns. In light of the fact that none of them comprises a foundation inscription, this is not surprising. The unique inscription at Avan (A.6), with its remarkable combination of Armenian and Greek words, interspersed with Armenian cryptograms, was carved onto the exterior of the eastern elevation.53 The solitary Armenian inscription found at Zvart'nocc(A.9) was carved onto a sundial which was unearthed in the course of excavation. The funerary inscription from Naxiavan/Noramanuk (A.13) was carved on a tombstone in one of two underground
4 The syntax of the inscription-and specifically the use of the instrumental U/,Jll/ U--suggests that his appointment was consequent to, and dependent upon, the foundation of the church. 5oJohn 3:29. 51Orbeli, Izbrannyetrudy,401-4; Strzygowski, Die Baukunst, 1:46-49. 52 This contention is examined below. K. Lafadaryan, Erevan. Mijnadaryan husarjannert' ev vimakan arjanagrut'yunnerW (Erevan, 1975), 190. The 5' cryptograms are based on the numerical values attached to each letter of the Armenian alphabet, exchanging those letters representing 1-9 with 1000-9000 and 10-90 with 100-900. This system resembles the Coptic "Era of Diocletian," used until long after his reign.

40

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

vaulted crypts.54 Finally, Uxtaytur's inscription (A.14), recording the provision of a water source, is located on two adjacent blocks on the southwestern pier inside the cathedral church at Tcalin.55 In summary, the dedicatory inscriptions at Bagaran (A.3), Alaman (A.4), Bagavan (A.5), Mren (A.7), Nax'avan (A.8), and T'alin (A.12) should be accepted as original by virtue of their location and layout alone, independently of any palaeographical comparison. Furthermore, there is no reason to question the authenticity of those inscriptions at Hi-ip'sime (A.2.1-2), Avan (A.6), Zvart'noc' (A.9), AruC (A.11), Naxciavan/Noramanuk (A. 13), and T'alin/Uxtaytur (A.14) on the basis of their location. It seems likely that at least three of the inscriptions at Mastara should be accepted as original (A.10.1-3), although the compressed nature of their layout challenges this proposition. Only the composite inscription at Tekor (A.1) has been shown to be a substitute or replacement inscription. The second test relies on palaeographical comparison. The results can be summarized succinctly. M. E. Stone reviewed the studies undertaken by G. Yovsip'ean and A. G. Abraho [Z], hamyan and traced the development of certain letter forms: ayb [11],za ['9], t'o [AI], and vew [4].56 In the course of this research, Stone accepted the originality of the inscriptions at Tekor (A. 1), Hiip'sime (A.2.1-2), Bagavan (A.5), Zvart'noc' (A.9), Mastara (A.10.1), Aru' (A.11), NaxEavan/Noramanuk (A.13), and T'alin/Uxtaytur (A.14). Previously, Yovsep'ean had accepted that the inscriptions at Mren (A.7) and T'alin (A.12) were genuine, including them in his album of Armenian palaeography. More recently K. Lafadaryan has argued that the letter forms of the Avan inscription (A.6) indicate a seventhcentury date.57Thus the majority of the inscriptions in Armenian within this corpus have been accepted by palaeographers as genuine. Furthermore, those that were not considered by these palaeographers-Bagaran, Alaman, and Naxiavan-all passed the first diagnostic text, being band inscriptions. When viewed collectively, the results of these two tests confirm the originality of eighteen Armenian inscriptions, found on fourteen structures. Unlike several of the Armenian inscriptions discussed above, the authenticity of the Greek and Arabic inscriptions found in Armenia has never been challenged. The inscription at Mastara (A.15) has traditionally been read as a simple date, referring to the twentyfirst year of King Peroz (A.D. 480). E V. Shelov-Kovediaev has argued, however, that the two words should be interpreted as names, Ifep6t(r(g) and (Ae)ovw6d(rfl).58 This revision sidesteps the issue of why a fifth-century Persian regnal year should be expressed in Greek on a mid-seventh-century structure and why Peroz is titled apXcov rather than Padottkag. The remaining Greek inscriptions included in this corpus are undated but all appear to originate in the early medieval period. The fragmentary inscription at Ereroyk' (A.16) has been dated on palaeographical grounds to not later than the seventh century.59Intriguingly, it was carved at the eastern edge of the southern fagade and so fits into the familiar
54 Atiian, Sirak, 136. In 1983 Thierry reported that this had been broken up by local villagers in the mistaken belief that it concealed buried treasure; see J.-M. Thierry, "A propos de quelques monuments chretiens du vilayet de Kars (III)," REArm 17 (1983): 353-57 and figs. 33-36. 1: microfiche 070, pl. C2. 55 Parsegian, ArmenianArchitecture, 56 3-11; Grv'ut'ean, 80-88; Stone, Sinai, 9-18. Yovs4p'ean, Abrahamyan, Hayoc'girev grc 'ut'yun, 57 Lafadaryan, Erevan, 190-91. 58 E V. Shelov-Kovediaev, "Zametki po grecheskoi epigrafike Armenii," IFZh (1986) 1: 66-68. arminienne, 45. 59Strzygowski, Die Baukunst, 1:31 and fig. 32; Khatchatrian, L'architecture

TIMOTHY GREENWOOD

41

pattern discerned above. The terse declaration of responsibility by Nerses for the construction of Zvart'noc' (A. 18) was described by Strzygowski as being located on the eastern side of the outer wall, although it is unclear whether this was its original position.60 Since several capitals from this church are decorated with monograms of Nerses, devised in "of Catholicos Narses," there Greek rather than Armenian and read as Kac0oXKo5oU Napaoou, can be little doubt that this Greek inscription is also original.6' It is significant, therefore, that whereas the church at Avan bears a single bilingual Armenian/Greek inscription from the time of the kat'olikos Ezr (A.6), the church at Zvartcnoc'constructed by Ezr's immediate successor, Nerses, originally carried Greek and Armenian inscriptions as well. The use of both languages was clearly deliberate and seems to reflect the political and doctrinal inclinations of both kat'olikoi; both are castigated for associating with the emperors Heraclius and Constans II and accepting the perceived heretical views of the Greek church.62 This issue will be discussed in greater detail below. Finally, the four inscriptions from Ejmiacin are located in a relief on the northern fagade of the cathedral.63 One is divided into two parts, flanking the lower register of a circular medallion (A.17.a and b). A third is located in a tabula ansata (A. 17.c), and the fourth (A.17.d) forms the medallion encircling a cross and ends within its bars.64Khatchatrian attributed them all to the fifth century, and there seems no reason to challenge his view.65 In contrast to the Greek inscriptions, two of the four relevant Arabic inscriptions contain specific hijra dates: A.H. 154 (770/771) and A.H. 170 (786/787).66 The two undated Arabic inscriptions are located immediately below the dated Arabic inscription at Zvartcnocc (A. 19). Given their close proximity to the first, Xa Catryan treated all three as a single inscription.67 However, in my view, three separate inscriptions can be distinguished. Not only does the second, middle inscription contain a separate, though incomplete hijra date; there are also differences in the style and inclination of the script as well as a large space between the second and third inscriptions. Dr. Robert Hoyland has examined these inscriptions from a palaeographical perspective and concluded that they all date from the eighth century. There does not appear to be any discernible pattern to the location of these were written onto the plaster surface of a column, Arabic inscriptions: those at Zvartcnocc indicating an internal rather than external location, whereas that at Aru' (A.20) was carved onto the exterior of the northern wall. However, their location is secondary when set against their very existence. The carving of Arabic inscriptions onto prominent Armenian churches in the second half of the eighth century raises a number of intriguing
Strzygowski, Die Baukunst, 1:30-31 and fig. 30. S. Der Nersessian, L'artarminien (Paris, 1977), fig. 30; Shelov-Kovediaev, "Zametki,"68-69. 62 T. W. Greenwood, "Sasanian Echoes and Apocalyptic Expectations: A Re-evaluation of the Armenian History Attributed to Sebeos," Le Musion 115 (2002): 3-4: 361-63 and nn. 150, 152, and 159. 63 7: microfiche 159, pl. B5. Parsegian, ArmenianArchitecture, 64 Although the relief contains inscriptions that could have been excluded on the grounds of relevance, collectively the four inscriptions merit inclusion. 65 Khatchatrian, L'architecture arminienne, 67-68. 66 XaC'atryan's corpus contains a second Arabic inscription from Zvart'noc'which purportedly bears the date A.H. 159; see A. A. XaC'atryan,KorpusArabskikh Armeni VIII-XVI vv. (Erevan, 1987), no. 7 and pl. v. nadpiesye Dr. R. Hoyland has reexamined the plate and reads the year as A.H. 257 (870/871). This inscription has therefore been excluded from this corpus. Following Hoyland, the purportedly 8th-century Arabic inscription from Bagavan (Xac'atryan, Korpus,no. 52 and pl. xxiii) contains the date A.H. 871 and not A.H. 171. 67 Xac'atryan, Korpus,no. 6 and pl. Iv.
61 60

42

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

questions about the ownership or control of these principal ecclesiastical centers as well as Arab administration and settlement. Finally, the mosaic Armenian inscriptions from Jerusalem have all been dated to the early medieval period-late fifth, sixth, or seventh centuries-on the basis of their content and the style of the accompanying representations and/or border decoration. Palaeographical analysis supports this broad contention. Two mosaics have attracted more specific comment. The inscription marking the resting place of the body of"Sugannan mother ofArtawan" (B.2) has been associated more precisely with the first half of the sixth century, on the basis that Artawan has been identified as the Armenian general Artabanes who occupied several key military commands under Justinian in Africa, Constantinople, and Thrace.68 On the balance of probabilities, this association seems less rather than more likely. Second, the mosaic commissioned by Ewstat' (B.6) has been dated on the evidence of a coin discovered in its mortar bed, tentatively identified as "Byzantino-Arab" and attributed to the mid-seventh century.69 Again there is no reason to question the early medieval dates attributed to these inscriptions.
LOCATION IN TIME

Eight of the eighteen Armenian inscriptions located in Armenia contain specific dates. These dates are expressed in several different ways. Collectively they repay close investigation because they chart important changes in the chronological systems employed in Armenia. They illustrate how contemporaries orientated themselves in time, and specifically in whose time they envisaged that they were living. At the same time, they invite comparison with chronological references in literary works dating, or purporting to date from, the same period. Before tracing these chronological developments, there is one characteristic common to all these dated inscriptions: the date is always placed at the beginning of the inscription. In isolation, this feature is of little consequence; however, in the context of late antique epigraphy in the Near East, it is very distinctive. The vast majority of inscriptions originating within the Byzantine Empire locate the date at the end of the inscription rather than at the beginning. Again, therefore, this corpus of Armenian inscriptions does not fit neatly into the mainstream Byzantine epigraphic tradition. The earliest surviving date occurs in the Bagaran inscription (A.3) and is calculated by reference to the regnal years of Xosrov II.70Thus the church was founded in the "thirtyfourth year of King Xosrov," and But Aruelean was murdered in Xosrov's thirty-eighth year.7 Although this dating mechanism is not found in any other inscription, it is used
68 Artabanes 2: PLRE 3A: 125-30. He was appointed magister militum per Africamin 546, magistermilitum praesentalis and comes foederatorum following his return to Constantinople in 546, offices he held until 549 when he was implicated in a plot to depose Justinian. However, within a year he had been restored to favor and was appointed magistermilitumper Thraciam,an office he occupied between 550 and 554, after which he disappears from view. Intriguingly he seems to have spent the duration of his final command in Sicily and Italy rather than in Thrace. It is instructive to note that Artabanes 1, who fought at the battle of Hippis in 549, was thus a direct contemporary of Artabanes 2. 69 M. E. Stone, "The New Armenian Inscriptions from Jerusalem," in Armenian Perspectives,ed. N. Awde (Richmond, 1997), 264. 70Although the short Greek inscription at Mastara (A.15) appears to supply a Persian regnal year from the 5th century, for the reasons outlined, this reading cannot be relied upon as secure. 71 Year 34: 19 June 622-18 June 623; year 38: 18 June 626-17 June 627.

TIMOTHY GREENWOOD

43

consistently in the History attributed to Sebeos. Of the twenty specific dates found in that text, no fewer than fourteen are based exclusively upon Persian regnal years.72Admittedly the precise wording of these dates is not consistent, implying that they derive from more than one underlying source, but it is clear that this was the preferred chronology in use in Armenia before 630.7"The same pattern emerges from those chapters in book 2 of the History of Atuank'which describe the devastation experienced by the region of Caucasian Albania between 624 and 630.74 These chapters contain four specific dates, all of which have been calculated by reference to Persian regnal years.75Finally, a solitary extant colophon attached to a miscellany copied in the thirteenth century states that Eznak translated this book in the "Twenty-ninth year of Xosrov, king of kings, son of Ormizd."76Consistent recourse to Persian regnal years in these primary sources before 628 is unsurprising, given that three-quarters of historic Armenia and the whole of Aluank' had been integrated into the Sasanian Empire for more than two centuries. However, the deposition and murder of Xosrov II in February 628 precipitated a period of sustained political instability within the Persian Empire, as a succession of rival claimants challenged one another for the throne and ruled, sometimes consecutively, sometimes concurrently, in different regions of the empire.77 This in turn provoked political and chronological chaos in Armenia, echoes of which can still be discerned within the Bagaran inscription (A.3). In the first place, the completion of the church is dated uniquely in terms of an event referred to previously in the inscription, namely, the death of But Airuelean; the text refers to the church being finished "three years" after his murder, rather than by reference to the regnal year of a Persian king. More importantly, the inscription locates this event in time by employing a synchronism based upon several Armenian figures; it refers to a time when Varaztiroccaspet was marzpan, when Vahan Aiielean was lord, and when an unknown figure, almost certainly a cleric and probably the local bishop, had responsibility for oversight or administration (the Armenian word hogabarjut'iwn implies an ecclesiastical authority). Nor is this an isolated example. The inscriptions at Alaman (A.4) and Mren (A.7), dating from 636/637 and probably 639/640 respectively, also employed synchronisms, and book 2, chapter 50 of the History of Atuank 'opens with a complex
72 Sebeos, Patmut'iwn Sebiosi, ed. G. V. Abgaryan (Erevan, 1979), 67, line 2 (Xosrov I); 94.11, 100.3, 104.6, 106.8, 111.11, 111.32, 112.9, 112.29, 112.31, 115.28, 121.5, and 123.15 (Xosrov II); 163.29 (Yazkert III); trans. and historical commentary by R. W. Thomson and J. D. Howard-Johnston, TheArmenianHistoryAttributed to Sebeos,2 vols., Translated Texts for Historians 31 (Liverpool, 1999). 73 Compare Sebeos, 111.32: "in the twenty-first year of king Xosrov," with Sebeos, 115.28: "on the 19th day in the month Margac', which was the 28th day of the month, in the 25th year of Apruez Xosrov." 74 Movsjs Katankatuaci: Patmut'(wn Atuanic'Asxarhi, ed. V. Arak'elyan (Erevan, 1983), 127.4-149.14; trans. C. J. E Dowsett, TheHistoryof the CaucasianAlbaniansbyMovsis Dasxuranc'i,London Oriental Series 8 (London, 1961), 75-92. Although the HistoryofAtuank'has been attributed to Movses Dasxuranc'i or Movses Kalankatuac'i, neither attribution is secure and it remains an anonymous composition. Atuanic, 128.10, 135.5, 140.17, and 142.8; Dowsett, CaucasianAlbanians, 76, 83, 86, and 87. 75 Patmut'iwn 76 A. S. Matcevosyan, Hayerenjeragreri (5-12 dd.) (Erevan, 1988), no. 19. This is similar to one hisatakaranner of the regnal formulae employed in Patmut'iwn Atuanic. "in the month Mehekan, in the 35th year of Xosrov Ormizdean, king of kings of Persia"; see Patmut'(wn At-uanic, 128.10; Dowsett, CaucasianAlbanians, 76. Ta'rikh al-rusul ed. M. J. de Goeje, 13 vols. (Leiden, 1879Sebeos, 129.22-130.34; al-Tabari, 77 wa-al-mul~tk, V The Sasanids,the 1901), 1:1045-67; trans. and commentary by C. E. Bosworth, TheHistoryof al-TabariVolume (New York, 1999), 381-410. A. Christensen, L'Iransousles Sassanides(CopenByzantines,the Lakmids,and Yemen hagen, 1944), 498-500; D. Sellwood, P. Whiting, and R. Williams, An Introductionto Sasanian Coins (London, 1985), 159-74; H. M. Malek, '"ASurvey of Research on Sasanian Numismatics," NC 153 (1993): 238-40. Cf. W. E. Kaegi, Heraclius, Emperorof Byzantium(Cambridge, 2003), 184-85, 202-3, 212-13, and 258-59.

44

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

synchronism defining a point in time between June 632/633 and January 638.78 This decade witnessed an important movement away from reliance upon Persian regnal years and toward individual synchronisms, devised through a combination of local and national figures, both secular and ecclesiastical. The selection of these figures was deliberate, reflecting the particular context in which the synchronism was created and thus explaining the lack of uniformity between them. As we shall see, the selection also intimates something of the hierarchy of power within the locality, as perceived by a contemporary. There is one striking omission from these synchronisms: the Armenian era dating system is not found in either of the historical texts specified above, nor in any of the seventh-century inscriptions. The emergence of the synchronism was not the only significant development in this decade. Armenian inscriptions began to employ the regnal years of Heraclius, both in isolation, as at Bagavan (A.5), and as a constituent part of a synchronism, as at Alaman (A.4) and Mren (A.7). Unlike the neutral references to Xosrov II discussed above, these three Armenian inscriptions accord Heraclius several very favorable epithets. Thus the Bagavan inscription describes Heraclius as "protected by God"; at Alaman he is /,qU4,,6, "victorious." and at Mren he is This positive assertion of pw,.hpbti,,, "pious," pmpha,,,lq, the virtues of Heraclius is also a characteristic found in the key Armenian historical works which report this period. Although Heraclius is initially described simply as king or emperor in the History attributed to Sebeos, he is subsequently lauded as hprUhjbp , I,,,,,I in and the so-called "Det hpfuSq*,, "blessed, God-loving, fortunate."79Moreover, AuutP L fense of Armenian Orthodoxy," addressed to his grandson Constans II, Heraclius is "the God-loving, God-worshiping, God-given, victorious, In the he fortunate Heraclius."80 is rJLnl,, beneficent, History of Atuank, tb,,pt1t ppPU"Flavius Heraclius the Christ-loving emperor."8' Even the problematic ,",,,I.ke Ituub"t~ii, L pbq rFmI opening of the History of Lewond presents Heraclius as m,,,1 IU.JllLbLUqUfaUWIfi "God-crowned and of the The pious king inscriptions Romans."82 uhfLUn,,,l confirm in general terms the accuracy of these literary references; they indicate therefore J,,,r1n,,g, that highly favorable epithets were applied to Heraclius at the time.
bptelt4c ,,[6ltu4tp,
Patmut'iwn Atuanic, 280.5-8: "In the reign of Heraclius and in the shattered kingship of the Persians of in the office of kat'olikos of Armenia of Ezr and in the office of sparapetof Mueel, in the lordship of Yazkert, Siwnik' of Grigor, in the office of bishop of Mat'usalay, and in the office of prince of Aluank' of Varaz Grigor, lord of Gardman .. ."The terminus post quemis supplied by the accession of Yazkert III, 16 June 632, and the terminusante quemis fixed by the date of the battle of Qadisiyya, 6 January 638, at which both the sparapet,Musel Mamikonean, and Grigor, ter or lord of Siwnik', perished. 79 Sebeos, 131.9-10. so Ibid., 152.6-7. 81 Patmut'iwn Atuanic, 128.14; Dowsett, CaucasianAlbanians, 76. Although the epithet 4t6ptarog, "Christloving," is not associated with Heraclius in any surviving regnal formula or inscription, an epitaph from Jerusalem dating from 592 has been reexamined by D. Feissel and read as follows: ... .a[otX(eita)de 537 et son application dans Maupt-Kiou To01t]X(o)X(pikcno)u Tougto': D. Feissel, "La rdforme chronologique ' l'pigraphie grecque. Annees de regne et dates consulaires de Justinien H6raclius," Ktima: Civilisationsde et de RomeAntiques 18 (1993): 186. Earlier, Cair Masp. I 67009.1 (a petition) applies the epl'Orient,de la Grace P. ithet tn6Xptoarog to the emperor. meci vardapetiHayoc4ed. K. Ezean (St. Petersburg, 1887), 3.8-9; trans. and 82 Lewond, Patmut'(wn Lewondeay of the Armenians (Wynnewood, Pa., commentary by Z. Arzoumanian, History of Lewond the Eminent Vardapet 48. 1982),
78

remembered as UMnLuJbUaU UUn.,,I,,,,,,,t, U, UM.nLLU,,JjaUM

r.,InP ,4lf

JWtJ,IPq, P4"Ljb1.,t, the valiant and

Lfl

TIMOTHY GREENWOOD

45

In my view, both the chronological mechanism and the positive epithets derive from the same source, namely, the regnal formulae of Heraclius employed in legal and official documents. These formulae have been preserved primarily in papyrus documents from Byzantine Egypt.83 Not only do they specify the date in terms of the regnal year of Heraclius rather than by indiction or other chronological system; they also contain identical is the exact Armenian epithets, albeit in Greek rather than Armenian. Thus Ur,UIp",4 is the Armenian of translation of equivalent OeobXacZKog,"protected by God"; rpwphLYUwl,, U is a direct translation of OEoomtre4;, fbL e6fop tao a og, "most pious"; and , ,, that "God-crowned."84 Furthermore it is,,L,,lnt possible literally "God-loving," IMui,1LMU-bU4p, the particular combination should be equated with "most holy."85 Unfortunately Oet6taog, of epithets found in the Armenian inscriptions is not duplicated in a single regnal formula. and However, both (in its superlative form e6Ei6;ei Uu,j,,4,/OEoU6aKwtog ,ptu,,m,/ in standard two of the found those formulae used in Byzanepithets comprise GoEpcraoaog) tine Egypt after 630. One final preliminary observation should be made. K. A. Worp calculated that "the number of post-630 documents from Egypt dated after Heraclius only and that of post-630 documents from Egypt dated after Heraclius and Heraclius Jr. is almost even."86Therefore, the dearth of any reference to Heraclius's son in the Armenian inscriptions or literary sources should not be construed as a matter for concern. The other epithets found in the historical texts do not appear in the surviving regnal formulae. There is, however, a second potential vehicle for the transfer of such epithets to Armenia, namely, the lengthy protocols found in the headings to imperial legislation and correspondence.87 Such protocols listed all the titles claimed by the emperor. The first noand vtfcrig; the Armenian equivvella of Heraclius includes the terms eUepig, Ei iru'ig, alents for these words are respectively "pious," 7Iq7iz,,4S, "fortunate," and pptwti,m,, two of Armenian these equivalents appear in the long list Significantly jW",,,l, ,"victor."88 of epithets applied to Heraclius in the draft Defense document, whereas the third, has undergone a minor change to Although no such protocols pFUP,,,t7,t, u,,in.,,r,"m. from the reign of Heraclius survive in Armenian, we know from elsewhere in the History attributed to Sebeos that a letter of congratulation was sent from the kat'olikos to Heraclius, when the latter was in Jerusalem to oversee the return of the True Cross on 21 March
R. S. Bagnall and K. A. Worp, "Regnal Formulas in Byzantine Egypt," BASP suppl. 2 (1979): 68-73; eidem, Chronological of ByzantineEgypt, 2d ed. (Leiden, 2004), 95-98, 267-71; K. A. Worp, "Regnal ForSystems mulas of the Emperor Heraclius,"Journal ofJuristic Papyrology 23 (1993): 217-32. 84 For the exact Greek equivalents, see Nor Baigirk 'HaykazeanLezui, ed. G. Awetik'ean, X. Siwrmelean, and M. Awgerean, 2 vols. (Venice, 1836, 1837; repr. Erevan, 1979, 1981). For the presence of IU1......anLUII,4/ "protected by God," in Egyptian papyri, see Worp, "Regnal Formulas," 227 (b); 228 (d); 230 (g); OeosO5acKog, and 231 (j); teMP/ o tazog,"pious"; ibid., 218 (1) and (2); 219 (3); 220 (4); 223 (12); 226 (a); 228 (c), rtp,tu=,,,,/e "God-crowned": ibid., 221 (7) (d), and (e); 229 (f) and (g); 230 (h) and (i); and , ,,,t UmL.... and (8); 222 (9); 223 (12) and (13); 226 (a); 228 (c); 229 (f); 230 t4bw/Oeore4~}g, (h) and (i); and 231 (j). It is intriguing to note that "protected by God," is consistently applied to the sons of Heraclius rather than U;,,qt,4/0eoO1Xcau-og, himself and that all the papyri containing this epithet postdate 630. For evidence that this epithet Heraclius was also applied to Heraclius, see the Smyrna inscriptions, note 93 below. 85 Worp, "Regnal Formulas," 218 (1); 221 (7); 222 (9); 223 (12) and (13); 228 (d); 229 (f) and (g); 231 (j).
83

86

Ibid., 225.

87J. Konidaris, "Die Novellen des Kaisers Herakleios," FM 5, Forschungen zur byzantinischen Rechtsgeschichte 8, ed. D. Simon (Frankfurt am Main, 1982), 33-106. 88 Konidaris, "Die Novellen," 62. This novel is usually dated 1 May 612 because it specifies the Kalends of May, indiction 15: ibid., 54 and 72.

46

A CORPUSOF EARLY MEDIEVAL ARMENIANINSCRIPTIONS

630.89 Moreover,irrespective of whether or not it was ever sent, the draft Defense docu-

ment attests written communication between Constantinople and Armenia in the early years of the reign of ConstansII. These Armenian inscriptionsreveal that there was a pattern of correspondencebetween the Byzantine Empireand Armenia during this decade, although it is not possible to identify the parties involved or the nature of that correspondence. The Armenian inscriptionsat Alaman (A.4), Bagavan(A.5),and Mren (A.7) therefore depend upon regnal formulae of Heraclius,both for chronological precision and for the laudatoryepithets applied to the emperor.The epithets do not of themselves reflect contemporaryArmenian perceptions of Heraclius,being integral to the formula from which the regnal year was taken. However,the translationand inclusion of such epithets in the text of an inscription implies tacit acceptance of them, revealing that Heraclius enjoyed a high reputation in Armeniaafter 630, as one might have anticipatedafter his startlingvictory over Xosrov II and the dramaticreconfiguringof the balance of power within Armenia which followed. Indeed, he is describedin the Defense document as q4,InpMh,'Ijnps, in Greek Tmrjpt686rg, a Christlikequalitythat finds expression in the projectionof Heraclius as the redeemer of the True Crossin the exterior relief at Mren.90 The link between regnal formulae of Heracliusand epigraphy is not confined to these Armenian inscriptions. In 1740 Kemplen copied a fragmentaryepitaph from a sarcophagus in Constantinople, now lost, whose last line Feissel has reconstructed as follows: toi K0 toi toiou io'l[ou~ tl'].91 Whether or not ETOXug 'HpaKXio- e6o-P(EodGtoiJ) (K)o) aUT[ou one accepts his reconstruction,the conjunction of e604eearGT0ov and of regnal yearsis exin similar 4 in the inscriptionat Althe Novella and conjunction tremelyinteresting,given TwoGreekinscriptionsfrom Smyrnasupply further insight.93 aman, albeit in Armenian.92 The first reads:
Kv < >tavtivmvvt6v vi~v 'HpaKrikou aQ 'HpaKXilow t6v e3vapFcyaozv cati 0*ko ig6l6v yrnt(o65v) '(oo)Kv

The second is fragmentary:

[zt6vviov Kvoz]vr]aivv

a v'Hpa[K-Xitox Iv Ec)[ [i X-ioZ6-v 0]oF0Xd-ktc i[g6 tot65v]

icai 'Hpa-]

The epithet "protectedby God,"common to both inscriptions,is also found in the Bagavan inscription (A.5). Therefore, the two epithets used in the first inscription at Smyrna
Sebeos, 118.8-17; Thomson and Howard-Johnston, Sebeos,2:210. Sebeos, 152.7. 91 Feissel, "La r6forme chronologique de 537," 186-87. "Year20 of the most pious Heraclius and year 18 of his son." Heraclius was enthroned on 5 October 610; his twentieth year therefore ran from 5 October 629 until 4 October 630. His son Heraclius Constantine was crowned co-emperor on 22 January 613; his eighteenth year began on 22 January 630. 92 Konidaris, "Die Novellen," 94. 93 Die von SmyrnaTeilII, 1, ed. G. Petzl, Inschriften Griechischer Stadte aus Kleinasien 24.1 (Bonn, Inschriften 1987), nos. 851 and 851A. These are translated: "Of Heraclius and Heraclius the new Constantines our most pious and God-protected emperors," and "Of Heraclius and Heraclius the new Constantines our Godprotected emperors, many years."
90 89

TIMOTHY GREENWOOD

47

occur in the Armenian inscriptions at Alaman (A.4) and Bagavan (A.5). This striking coincidence may be explained by adopting the same interpretation in respect of the Smyrna inscriptions, namely, that these too reflect the content and style of imperial regnal formulae. In other words, inscriptions both inside the Byzantine Empire and in Armenia depended upon contemporary imperial regnal formulae for chronological precision. In the course of exploiting these formulae, the laudatory epithets applied to the emperor were also transferred, both in Greek and in Armenian translation. In contrast to the several inscriptions associated with the 630s, only one Armenian inscription can be linked to the following decade. The first Mastara inscription (A.10.1) states that the church was built in the years of T'eodoros, bishop of Gnunik'. One Lord T'codoros bishop of Gnunik' is recorded attending the Council of Dvin, convened "at the command of the blessed katcolikos of Armenia Nerses Sinot and other bishops" in the "fourth year of Emperor Costandianos, pious [barepavt]king of the Romans."94These canons confirm that this see was occupied by a bishop T'eodoros in the mid-640s, and the church has been dated on the strength of this evidence. This inscription is unique because it represents the only known instance of the tenure of the local bishop being used as the solitary dating mechanism. Tenure of local episcopal office is, however, one of the constituents in the synchronisms used at Bagaran (A.3), Alaman (A.4), and Mren (A.7)-and in the synchronism found in book 2, chapter 50 of the History of AtuankI.5 This chapter gives additional insight into this dating mechanism. Its self-proclaimed author, one Yovs4p, displays a precise knowledge of the length in office of two bishops of ArccaxMihr and his successor, Andr&as.96 This implies the existence of records listing successive and the duration of their tenure, from which a local dating mechanism could be bishops devised. Although such local records might appear conjectural, one such list has been preserved in a colophon, dated 155 of the Armenian era (706), of MsJerusalem 849.97 The colophon contains a list of the successive bishops of Ars arunik' from Kndak down to Grigoris, supplying the tenure of each in years:

94 Kanonagirk'Hayoc,ed. V. Hakobyan, 2 vols. (Erevan, 1964, 1971), 2:200. This preamble confirms the continuing influence of Byzantine regnal formulae upon contemporary Armenian chronological practice. It defines the year exclusively by reference to the reign of Constans II and qualifies his royal title with the familiar epithet "pious."Although the History attributed to Sebeos contains seven dates calculated by reference to the regnal years of Constans II, none of these affords him an epithet, laudatory or otherwise: see Sebeos, 141.10, 142.16, 146.11, 163.29, 164.12, 169.18, and 170.19. A colophon appended to a mathematical text records that "I Anania Sirakac'i from the village of Anenic'" began the undertaking "when God the Word had filled 668 years from the Virgin's womb, and in the eleventh rotation, in the 19th year of Constans": Mat'evosyan, no. 24. In this instance a synchronism has been attempted using the birth of Christ, an indicHiMatakaranner, tion date, and an imperial regnal year. Although the figures do not correspond to a single year, the combination of these Byzantine chronological mechanisms is significant, indicating that Byzantine practice continued to be employed in Armenia into the 660s. The Aruc inscription (A. 11) supports this proposition. 95 At Bagaran (A.3), the identity of the bishop and his see are unknown; at Alaman (A.4) and Mren (A.7), the bishop is named as T'eo(v)p'ilos, bishop of Alarunik'. Intriguingly the Canons of Dvin (September 644/645) identify one Grigor as bishop of Ariarunik' and one T'eop'ilos as the bishop of Aimunik': see Hakobyan, Kanonagirk',2:214. It is possible that the two have been confused. See above, note 78, for a full translation of the synchronism in Patmut'iwn Atuanic. 96 Patmut'iwn Albanians,181. Mihr: twelve years; Andreas: eleven years. Aluanic, 280.16-19; Dowsett, Caucasian 12 dar) (Ant'ilias, 1951), no. 17. According to this list, daric'min(cew 97 G. Yovsep'yan, Yifatakarank'jeiragrac'(5 T'Vop'ilos was succeeded by Movsis, not Grigor.

48

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS Bagrewandand Ariarunik'were establishedby Saint Grigor as the diocese of one bishop
the blessed bishop Movses. Therefore, until the death of the Father of poets [k'ert'otahayr],

and the house of Kamsarakan, the bishop was first: having separatedArvarunk' 46 Kndak, years;Yohan, 11; T'atul, 9; T'ovma,40; Havukn, 16; T'ovma, 11; Movses, 10; Yoys, 1; T'eop'ilos,26; Movses, 12; Lazar,27; Grigoris,33. Until 155 of the Era of this volume, it was this [year]in which GagikKams[a]rakan conseand news arrivedthat Su'an Kams[a]rakan the daughcrated the church of Vardanakert ter of Vahanpatrikios had inherited the name of faith confessorfor the sake of ChristGod, been in Xai-anof Mesopotamia.Grigorisstill another 7 years. tortured having In this context, it is also worth recalling that a list of Armenian kat'olikoi was exploited in the compilation of the History attributed to Sebeos, and similar lists have also been discerned within, or appended to, the Narratio de RebusArmeniae,the HistoryofAtuank'(a list of kat'olikoi of Aluank' rather than Armenia), and the History of Yovhannes DrasxanaThere seems to be little doubt that such records were maintained, at both a local kertcci."O and a metropolitan level. The inscription at Mastara (A.10.1) corroborates this practice and illustrates how such records could be used for the purposes of dating local monuments.

Only one inscription is dated by reference to the reign of Constans II, namely,the famous panel inscriptionatAruE(A.11).The opening line specifiesthe 29th year of Constans. Unlike the earlier inscriptions,the year is representedin numerals rather than words,but this new practice does not seem to be significantwhen considered in the light of evidence from contemporarytexts. The History attributedto Sebeos contains five dates expressed Instead, it is the figure itself that has prompted considerable disquiet. using numerals.99 The traditional chronology of Byzantine emperors maintains that Constans II was murdered in his twenty-seventhyear.'00 This evident discrepancyled both Orbeli and Strzyof the to the gowski query authenticity inscription.1'0 How is this to be explained? In the first place, this inscription is not the only Armenian source to accord Constans II a longer reign than is generally accepted. The History of
Atuank'states that Constans II was assassinated in his thirtieth year.102 Setting to one side for the moment the accuracy or otherwise of these dates, they imply that the annual cycle in both Armenia and Aluank' continued to be calculated by reference to the regnal years of Constans II throughout the 660s, long after the emperor's ambitious plans for an exGreenwood, "Sebeos," 360-63; G. Garitte, La Narratio de RebusArmeniae, CSCO 132, Subs. 4 (Louvain, 1952), 401-45; Patmut'iwnAtuanic, 341.21-347.4 and Dowsett, Caucasian Albanians, 227-31; Maksoudian, 43. A separate list of kat'olikoi, attributed to Yovhannes Drasxanakertc'i and comYovhanntsDrasxanakertc'i, posed in 897, was published as an appendix to the History of Samuel of Ani: Samuli k 'ahanayiAnec'woy Hawak 'munk ed. A. Ter-Mik'elean (Valariapat, 1893), 269-77. 'i groc'patmagrac, 99Sebeos, 104.6, 112.29, 115.28, 121.5, and 123.15. All refer to regnal years of Xosrov II; thus none of the regnal years of Constans II employ this abbreviation. 100Theophanes places this event in A.M.6160 (A.D.667/668) and equates it with the twenty-seventh year of Constans II: Theophanes the Confessor, Chronographia, ed. C. de Boor, 2 vols. (Leipzig, 1883-85), 1:351.14352.9; trans. and commentary by C. Mango and R. Scott, The Chronicleof TheophanesConfessor: Byzantineand Near EasternHistoryA.D. 284-813 (Oxford, 1997), 490-9 1. Nikephoros reports the length of the reign of Constans II as twenty-seven years: Nikephoros,Patriarch of Constantinople.Short History, ed. and trans. C. Mango, DOT 10 (Washington, D.C., 1990), para. 33. 101 Orbeli, Izbrannyetrudy,401-4; Strzygowski, Die Baukunst, 1:46-49. Patmut(iwn Atuanic, 196.19: "in the thirtieth year of the kingship of Constans, eunuchs had undertaken 102 to murder him."
98

TIMOTHY GREENWOOD

49

tended network of Armenian, Albanian, and Persian clients across the Caucasus to the Caspian and beyond had been thwarted by Mucawiyaand Constans II had turned his attention instead to the strategic potential of Greece, Italy, and Sicily. The significance of this continuity is harder to determine. The use of a Byzantine chronology may reflect the impact Constans II had had upon Armenia and Aluank' in the 650s, ongoing contact by imperial authorities during the 660s, the dearth of an alternative, accessible chronology, or a combination of any of these. From a Byzantine perspective, however, the single issue is their challenge to the traditional date for the murder of Constans II. In this respect they are supported by a source compiled in Italy. The LiberPontificalisstipulates that Constans II was murdered in his bath "on the fifteenth day ofJuly in the 12th indiction."'03 This establishes the year 669. Therefore, a source composed in Rome accords Constans II a reign of twenty-seven years and ten months; another way of presenting this information is to say that Constans came within two months of entering his twenty-ninth year. Given the proximity between the place of composition of the LiberPontificalis,Rome, and the site of the emperor's murder, Sicily, it is strange that this evidence has not been preferred to the date supplied by Theophanes, whose chronological imperfections are well known."04 When these fragments of evidence from both the eastern and western peripheries of the Byzantine world are combined, they support the proposition that Constans II was murdered on 15 July 669, at the very end of his twenty-eighth year. In any event, even if one rejects this attempt to redate the assassination of Constans II, this is not a sufficient reason to doubt the authenticity of the inscription at AruE. In the light of the confused imperial succession after the murder of Constans in distant Sicily, and the apparent lack of another annual mechanism in Armenia, the continued use of the regnal years of Constans II appears eminently sensible. The parallel with the retention of Yazkert III on coins minted long after his death has already been made. The Aru' inscription (A. 11) is the latest Armenian inscription of the seventh century to refer to a Byzantine emperor. There is then a gulf of some three hundred and fifty years before an emperor, Michael IV, is mentioned again in an inscription.105 This reference needs to be distinguished from the other examples because it does not occur in the context of a dating mechanism. The Aru' inscription, however, is not quite the end of the trail. The Georgian band inscription located on the exterior of the church of Sion at Samsvilde
L. Duchesne, Le LiberPontificalis:Texte,introductionet commentaire, 2d ed. with amendments by C. Vogel, 3 vols. (Paris, 1955-57), 1:344 and n. 7: "Et postmodum, XV die mensis iulii per XII indictionem praedictus imperator in balneo occisus est"; trans. R. Davis, TheBookof Pontiffs (LiberPontificalis),Translated Texts for Historians 5 (Liverpool, 1989), 72. Davis argues that from the time of Honorius (27 October 625-12 October 638), "there is no real doubt that we are dealing with a series of contemporary additions." P Grierson, "The Tombs and Obits of the Byzantine Emperors (337-1042)," DOP 16 (1962): 49, states unequivocally: "The date of Constans' murder in Syracuse can be fixed as 15 July 668. The date is given by no known Greek source, but the Vita Vitalianiin the LiberPontificalisdates it 15 July of the second (sic) indiction, i.e. 669, and though the last is evidently a slip-it must have been the first (sic) indiction-the author is so well-informed about Constans' doings in the west that 15 July can hardly be an error."It seems strange to accept the day of the month stipulated in this text yet reject the indiction date out of hand. 104 Mango and Scott, Theophanes,lxvi-lxvii: "The historian would be well advised to subject all of Theophanes' dates to careful scrutiny." 105 DHV 1: no. 129 and pl. xv. The inscription begins: "In 484 of this era I Ablarip marzpantook a message from Smpat Sahanfah to the emperor of the Greeks Mixel in Constantinople .. ."This equates to the year 13 March 1035-11 March 1036.
103

50

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

(Samswlde) attests the use of this means of dating in the third quarter of the eighth century.106 It records that construction of the church began in year 20 of the reign of Constantine and was completed in year 3 of King Leo. The twentieth year of Constantine V ran from 18 July 760, and the third year of Leo IV began on 14 September 777. The use of this chronological mechanism implies a pattern of contact and communication between the Byzantine Empire and successive eristavis, Varazbakur and lovane, who were related in some way to the pitiax? of Gogarene.'07 This discourse is not otherwise attested. The inscription indicates that as late as the third quarter of the eighth century, Byzantium maintained contact with princely families whose territorial possessions lay far beyond imperial borders. It is tempting to envisage a similar engagement with Armenian princely families at this time, but this must remain conjectural.108 The Aru' inscription is also the last dated Armenian inscription before the late eighthcentury inscription of Uxtaytur at T'alin (A. 14). This employs an entirely new mechanism, namely, the Armenian era. It states that it was carved in the year 232 (15 May 783/13 May 784) and thus constitutes the earliest extant inscription to use this chronology. This has great importance for the study of the Armenian era system because it supplies a secure terminus ante quem. As noted previously, the earliest original manuscript to contain an Armenian era date is the famous Gospel of Queen Mlk'.1?09The first use of the Armenian era by an Armenian historian occurs in the History of Lewond.10oThis text contains a single Armenian era date, 234 (14 May 785/13 May 786), provided one accepts the correction supplied by the History of Step'anos Taronec'i.III However, the date of composition of Lewond's History remains contentious; if a late ninth-century date is preferred, the significance of this Armenian era date is diminished.112 The earliest Armenian era date in the
I am very grateful to Professor R. W. Thomson for supplying the following translation, based on the reconstruction proposed by L. Muskhelishvili and reproduced by N. Chubinashvili; see note 32 above. The addition material in square brackets was proposed by Muskhelishvili, but the inscription seems to run on satisfactorily without this insertion. "O Christ, all-holy born of God, have mercy on the builders of this holy church, Varazbakur, lovane and ... by origin pitiaxss. In the year 20 of the reign of Constantine the foundation was laid by the hand of the mamp'aliand eristaviVarazbakur.And the building was not yet ready (when) Varazbakur fell asleep and his last brother, the eristavi lovane, was zealous. Iovane completed ... the final construction of this holy church and its stoas. And in the year 3 of king Leon, on the day of dedication/enkenia, the consecration took place for six days with a great crowd." Both stoa and enkeniaare Greek words in Georgian script; their presence strengthens the proposed Byzantine connection. and 107 C. Toumanoff, Studies in Christian Caucasian History (Washington, D.C., 1963), 185-192, 260-64, 399-408. Both the sequence and chronology of the pitiax's of Gogarene remain contentious. Neither Varazbakur nor lovane register in these studies. 108 The History of Lewond supplies a vivid account of the uprisings in Armenia in the early 770s and their suppression: Lewond 135.6-154.17; Arzoumanian, History of Lewond, 127-40. These are traditionally presented as a local Armenian reaction to excessive taxation. If one accepts the contention that there was contemporary correspondence between the Byzantine emperor and both Iberian and Armenian princes, there is an argument that the uprising was inspired, or at the very least encouraged, by the empire. 109 See above, note 1. 110 Lewond, 165.23-25; Arzoumanian, Historyof Lewond, 147 and n. 7. The date given in the text is 233 (14 May 784/13 May 785). This date is then synchronized with the brief 111 caliphate of Mfisa al-Hadi (4 August 785-14 or 15 September 786), which in fact corresponds to the years 234 and 235. However, Step'anos Taronec'i extracted material from the History of Lewond, including this passage, and his History records the year as 234: Step'anos Taronec'i, Step'anosiTaronec'woy Patmut'iwnTiezerakan,ed. S. Malxazeanc' (St. Petersburg, 1885), 133.23. The letters q- and 9. are easily confused. 112 The contention that the History of Lewond reflects a late 9th-century perspective will be advanced in a
106

TIMOTHY GREENWOOD

51

History of T'ovma Arcruni is 300 (28 April 851/26 April 852)."11 As noted previously, Yovhannes Drasxanakertc'i's History is extremely sparing in its use of dates; its earliest Armenian era date is 302 (27 April 853/26 April 854)."14 Although book 3 of the Historyof Atuank 'dates events from the seventh century by reference to the Armenian era, it is clear that the underlying chronicle source from which these passages were taken was compiled long after these events."15 Evidently the author of this chronicle, active at the start of the tenth century, had resolved to calibrate events by reference to the Armenian era, even if this resulted in serious chronological error. Thus Uxtaytur's inscription has a particular importance, confirming that the Armenian era was in use by 783/784, seventy years before the earliest reliable reference in an Armenian historical text. In relation to this chronological mechanism, there is one other category of material to consider. Several Armenian colophons contain early Armenian era dates. The earliest is that attached to the original Armenian translation of Socrates' EcclesiasticalHistory (SI): In the six thousandtwo hundred and fourthyear of the transitory existence[6204 - 5508 = as the of Saint the senior church of the universal motherhas, 696], chronography Sop'i and seven the hundred and fourth of the city Constantinople, coming of Christ[704 - 9 = = 6 June 695/5 June 696], in and one hundred and + 144 of the era 695] [551 forty-four in the ninth indiction [1 Sept. 695/31 August 696], in the first year of the reign of the second Leo, the autocrator Augustos [695/696], P'ilon T/Sirakac'i translated through an original undertaking these books, a special abridgement from the Ecclesiastical History of Sokrates, having omitted the personal remarks and letters of others and discourses because of a shortage of paper and parchment that existed in foreign parts.

O Lord Nerseh apohypaton patrikios, you who are a builder of churches, may you always have peace from Christ, who is the giver of peace, may you constantly have the zeal of pious kings to sustain the labors of piety in yourself, and never having occasion may you wander from the paths of righteousness. And may you not be drawn and led astray by opponents who will fall into ruin, but contemplating what is above, may you evade assassins and pursue righteousness, faith, love, striving after eternal life, [p UF,,.cIm/ul i,",npry] like the pious king T'eodos. And may you prosper in everything that you do and temptation be defeated through prayer which you have heard from this History and which we have seen. For the peace of Christ shall dwell in you abundantly and in your churches and country. Because you desired the translation of this History, with the same you shall walk the path of kings, through the grace and benevolence of our Lord and Savior Jesus Christ, to whom with the Father Almighty and the Holy Spirit, glory, power, and honor, now and always and for ever and ever, Amen, Amen.116 forthcomingarticle.Centralto the argumentis a propheticstatementin the text: "Beholdthe time of your salvation has come, for now the scepterof kingshipshallreturnvery soon to the house of T'orgom.. ."(Lewond, 141.18-21). This prophecyis most easilyinterpretednot as the predictionof a contemporary8th-centuryauthor but rather as the confident expressionof one who had witnessedthe reality of a revivedkingdom of Armenia through the coronationof Asot I Bagratunion 26 August 884.
113

facsimilereproduction,with new introductionby R. W.Thomson (Delmar, N.Y., 1991), 106.16-17; trans.and

Artsruni:Historyof the House of theArtsrunik' (Detroit, 1985), 173. commentary by R. W. Thomson, Thomas 114 Yovhannes Drasxanakertc'i, 124.27-125.1; Maksoudian,Yovhannes Drasxanakertci,25.69. 115 Patmut'iwnAtuanic, 316.16-338.12; Dowsett, Caucasian Albanians, 206-24. The most serious, and in-

Thomas Arcruni, Patmut'iwnTannArcruneac,ed. K'. Patkanean (St. Petersburg, 1887; repr. Tiflis, 1917),

triguing, error is the transpositionof the Arabconquestsfrom the reign of Heracliusto that of his son Constantine,who is accordeda reign of three yearsratherthan 120 days. 116 no. 28. Mat'evosyan, Hi?atakaranner,

52

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

As R. W. Thomson has noted, there are problems in accepting the authenticity of this colophon.'11 Although it is attached to SI, it describes the shorter, later adaptation of this work (SII). SII is a reworking of SI and not a fresh translation from a Greek original. Even the reference to paper, k'artiz,is problematic because this did not circulate in the Near East until the middle of the eighth century. Furthermore, there are four separate, authentic colophons which record the translation of Greek patristic texts into Armenian in Constantinople between 713 and 717.118 Dawit', hypatosand kinai of the royal table, and Step'annos, ot, were responsible for these translations. Although these four colophons priest and k 'ert contain synchronisms devised in much the same way as that employed by P'ilon, none of them uses an Armenian era date. It is worth remembering, however, that the five elements of Pcilon'ssynchronism, including the Armenian era date, correlate very neatly with one another.119 There seems little doubt that the synchronism evinces a Constantinopolitan origin. It refers to the church of St. Sophia and stipulates both an imperial regnal year and an indiction date. Nor does this colophon constitute the only contemporary evidence for St. Sophia as a repository of chronological knowledge. It has been suggested that the ChroniconPaschale, compiled in the third decade of the seventh century, was composed by a member of the clergy of St. Sophia under the patronage of Patriarch Sergius.120 Although Thomson has rightly doubted the association between the colophon and SII, the synchronism, including the Armenian era date within it, may be genuine. Four separate, unrelated colophons from the first half of the eighth century employ the Armenian era system as their exclusive chronological mechanism. The list of the bishops of Ar'arunik' quoted above is dated 155 of the era (3 June 706/2 June 707).121 A history of the Holy Lance and other Mysteries of Christ was copied and translated "from Roman writing" into Armenian in Era 176 (29 May 727/27 May 728)."12 The colophon attached to Artavazd Abraham's account of the martyrdom of Vahan Goltcn notes that Armenia began to drink the cup of punishment in "one hundred and fifty-two of the Armenian era" (4 June 703/2 June 704) and stayed "in captivity until one hundred and eighty-six of the era (26 May 737/25 May 738), in which the martyrdom of Vahan occurred."'23 Finally, a famous collection of homilies was assembled and organized into an annual cycle by Solomon, leader of the community of Mak'enoc', in Era 196 (24 May 747/
117R. W. Thomson, TheArmenianAdaptationof the EcclesiasticalHistory of SocratesScholasticus,Hebrew University Armenian Studies 3 (Leuven, 2001), 9-13. nos. 31-34; see also Yovsip'yan, Yifatakarank'jei 118 Mat'evosyan, Higatakranner, agrac, nos. 18-20 for three of the four. These are attached to translations of books attributed to Dionysius the Areopagite, commentaries of Cyril of Alexandria, a commentary on Leviticus, and a miscellany of works by Gregory of Nyssa. The patristic focus of this translation initiative in Constantinople is significant because it supports the statement, repeated in each colophon, that the same two individuals were responsible for all four works. 119 Although the Christian era date is wrong by nine years, P Peeters maintained that "Lann&e 704 de l'Incarnation est ici comptde selon la pratique des chronologistes armeniens, qui partent de l'an -9. Correction faite, nous retombons sur l'annde 695." See P. Peeters, "A propos de la version armtnienne de l'historien d'histoire et dephilologieorientales,1, SubsHag 27 (Brussels, 1951), xix, 314. Socrate," in Recherches 120 M. and M. Whitby, ChroniconPaschale284-628 AD,Translated Texts for Historians 7 (Liverpool, 1989), xxii-xxviii. 121 Yovsep'yan, Yifatakarank 'jeiagrac, no. 17. 122 no. 36. Mat'evosyan, Higatakaranner, 123 Ibid., no. 38. For further discussion of this colophon, see below, and note 223.

TIMOTHY GREENWOOD

53

22 May 748).124 Collectively these colophons attest the regular use of the Armenian era in the first half of the eighth century. Of course, the dearth of evidence precludes any judgment as to how widely this chronological system was used within Armenia; was it restricted to monastic communities or to certain regions or to particular documents? Nevertheless, a study of the literary and epigraphic sources suggests that the Armenian era was devised at some point after the death of Constans II and that it was in use in the first quarter of the eighth century.125 This corpus presents one final chronological challenge. The inscriptions at Bagaran (A.3), Bagavan (A.5), Mastara (A. 10.4), and Aru' (A. 11) all specify the month and day of the month on which the construction of the church was completed; the Bagavan inscription also gives the day on which construction began.126 Moreover, the Susannan mosaic inscription from Jerusalem concludes with the specific date Hori 18 (B.2). Although the exact combination of elements varies within these five inscriptions, they all adopt the same distinctive sequence: year, month, day of the month, and then day of the week. In so doing, they corroborate the structure of the three specific dates found in the History attributed to Sebeos. At the same time, however, these three dates may be distinguished. Far from commemorating the foundation of a church, they each define the final capitulation of a besieged center: Jerusalem, Dvin, and Arcap' k'.'27 How are these two groups of specific dates to be interpreted? As proposed previously, it seems most likely that the day on which a church was completed was recorded in its foundation inscription because it set the date for the annual commemoration of the benefactors of that church. Not only did this have a practical advantage in that the date was preserved on the structure itself for as long as it stood; it also had an obvious Old Testament analogue. 1 Kings 6 specifies when Solomon's Temple in Jerusalem was started: "In the four hundred and eightieth year after the Israelites had come out of Egypt, in the fourth year of Solomon's reign over Israel, in the month Ziv, the second month, he began to build The chapter also identifies when the Temple was completed: the Temple of the Lord."'128 "In the fourth year of Solomon's reign, the foundation of the house of the Lord was laid, in the month Ziv; and in the eleventh year, in the month of Bul, which is the eighth month, the house was finished."129 Moreover, in recording Solomon's dedication of the Temple, 1 Kings 8 identifies the exact date on which this occurred.'13 Thus the chronological precision displayed in the account of the construction of the Temple is replicated in the
124
125

Ibid., no. 39.

Peeters argued that the Armenian era system was devised in the 670s; see Peeters, "La version armenienne de l'historien Socrate," 317-19. 126 Bagaran (A.3): three years after the death of Lord But A-uelean, month Trn, day twenty. Bagavan (A.5): twenty-first year of Heraclius, month Hrotic', day thirty for start of construction and twenty-ninth year of Heraclius, month Navasard, day twenty, a Friday, for completion. Mastara (A.10.4): Of the month Arac', [day] 14. Aruci (A. 11): 29th year of Constans, month Mareri, day 15. 127 Capture of Jerusalem (Sebeos, 115.28-29): "on the 19th day [of the siege of Jerusalem], in the month Margac', which was the 28th day of the month in the 25th year of the kingship of ApruEz Xosrov." Capture of Dvin (Sebeos, 138.20): "the 20th day of the month Tre, a Friday."Capture of Arcap'k' (Sebeos, 146.11-12: "in the second year of Constans, in the month Hoii, on the 23rd day of the month, a Sunday at dawn." 128 1 Kings6:1. 129 1 Kings 6:37-38. 130 1 Kings 8:2-3.

54

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

Armenian dedicatory inscriptions. In my view, this echo is deliberate. Although this proposition might at first sight appear unlikely, a passage from the HistoryofAtuank' illustrates the influence of these chapters from 1 Kings.1'' The passage comprises a prayer of dedication purportedly recited by juanisr, prince of Aluank in the third quarter of the seventh century, on the occasion of the consecration of a church founded by him. Juanglr's prayer is examined in the following section, but for the present it is worth noting how closely it is modeled upon Solomon's dedication found in 1 Kings 8. This extends to recording the precise day on which the church was consecrated; we are told that Juansir's church was consecrated on the eleventh day of Arac'.132 The second group of specific dates should also be interpreted in terms of an Old Testament paradigm. 2 Kings 25 records the siege of Jerusalem by the forces of the king of Babylon, Nebuchadnezzar, and its eventual capture.'33 It records that Nebuchadnezzar advanced with all his forces to the walls ofJerusalem in the "ninth year of the reign of King Zedekiah, in the tenth month, on the tenth day of the month," and that the city fell in the fourth month of the eleventh year of Zedekiah, on the ninth day. A similar precision informs the description of the fall of Jerusalem to the Persians included in the History attributed to Sebeos. It records that the city fell "in the month Margac', which was the 28th day of the month in the 25th year of the reign of Apruez Xosrov." Although the two accounts differ in matters of detail, their scrupulous concern for chronological accuracy implies a relationship between the two. The accounts of the capture of Dvin and Arcap'k by Arab raiders were similarly influenced, being refracted through the prism of the Old Testament description of that first fall of Jerusalem. These centers were both under the control of the people of God and were captured by enemy forces after a siege. In summary, the literary and epigraphic sources confirm that a high degree of chronological precision was capable of being achieved in Armenia during the seventh century. They also reveal, however, that such precision was reserved for particular circumstances. Two such contexts have been defined, namely, the annual service of commemoration for the benefactors of a church and the fall of a besieged city or fortress. The former secured the date for the annual celebration of the church's foundation, and the latter mirrored the Old Testament account of the fall of Jerusalem.
DEDICATION AND PURPOSE

Fourteen of the eighteen Armenian inscriptions found in Armenia, together with one of the inscriptions in Greek, are associated in some way with the original foundation of a church. These dedicatory inscriptions tend to include two important details: the date of the foundation and the identity of the person or persons responsible for the foundation. A study of the second element reveals that these inscriptions can be divided into two groups: those recording the construction of a church at the instruction of a cleric and those attesting the initiative of a secular figure. The first group includes the foundations at Tekor (A. 1), Hi-ip'sime (A.2.1-2), Bagavan (A.5), Mastara (A. 10.1-4), and Zvart'noc' (A.18). Although the last lacks a full dedication in Armenian, we have already noted that the terse
131Patmut'iwn
133

132Patmut'iwn

Atuanic, 188.6-189.20; Dowsett, CaucasianAlbanians, 121-22. Atuanic, 187.21; Dowsett, CaucasianAlbanians, 121.

2 Kings 25:1-4.

TIMOTHY GREENWOOD

55

inscription in Greek and the Greek monograms carved onto several large capitals refer to the kat'olikos Nerses III. Independent notices in the Histories attributed to Sebeos and Yovhannes Drasxanakertc'i and the History of Atuank'all credit Nerses III with responsibility for the construction.'34 The second group of dedicatory inscriptions, attesting secular commissions, comprises those at Tekor (A. 1), Bagaran (A.3), Alaman (A.4), Mren (A.7), Nax'avan (A.8), Aruc (A.11), and T'alin (A.12). Tekor appears in both groups because the inscription is composite, reflecting two separate foundations; one part identifies Yohan kat'olikos, and the other names Sahak Kamsarakan. Arguably, this division into lay and clerical highlights a further distinction. Those inscriptions naming secular lords as responsible tend to justify the foundation in terms of intercession for themselves and their immediate family. The benefit to the individual is stated explicitly. Thus the construction of a church was recognized as a means of gaining remission for sins and escaping divine punishment. Those linked exclusively to clerics tend to be more reticent when it comes to articulating the spiritual advantages for the founder. Thus the inscriptions at Hiip'sime (A.2.2) and Zvart'noc' (A. 18) contain a brief invocation to Christ to "remember" the founder, and those at Tekor (A.1) and Bagavan (A.5) do not give any reason for the foundation, simply naming the individual concerned without further comment. There are, however, exceptions to both categories. Three of the inscriptions at Mastara refer explicitly to the anticipated redemption and expiation of the church's founder, the unworthy cleric Grigoras (A. 10.1-3). This mirrors the inscriptions from secular foundations. At the same time, however, two of the inscriptions assert that the church has a role as "a refuge for Mazdara" (i.e., Mastara) (A.10.2-3), the second adding that, through the intercession of St. Yovhannes, Mastara will prosper. These inscriptions attest an intriguing fusion of individual and collective benefit. Equally the church at Mren was built for the intercession of "the Kamsarakank' and Mren and all [?]."This asserts a much broader benefit, one that extended beyond the prince and his immediate family to his kin and, it seems, to the local community generally. The broad nature of this intercession is exceptional in the context of a secular foundation. There is no obvious explanation for this apparent distinction between ecclesiastical and secular foundations. It is possible that regular intercession for a clerical founder was somehow implied through his involvement in the church's construction and thus did not need to be expressed in the same way, but this is conjectural. The key observation is that both Armenian clerics and princes founded churches at this time. A similar impression is obtained from the written sources. Although the History attributed to Sebeos records the construction of the church at Avan by the so-called anti-kat'olikos Yovhannes and the building of the martyrium of Hrip'sime by the kat'olikos Komitas, it does not fail to acknowledge the central role played by Smbat Bagratuni in the rebuilding of the cathedral Thus the categorization identified from the dedicatory inchurch of St. Grigor in Dvin.'"35 scriptions holds true for literary sources as well: both clerics and princes were responsible for church foundations. Before advancing to consider those inscriptions that record something other than the original foundation of the church, let us return to the account of the construction,
Sebeos, 147.21-31; Patmut'iwnAtuanic, 236.9-10; Yovhannes Drasxanakertc'i, 88.13-18; Maksoudian, 19.47. Yovhannes Drasxanakertc'i, 135 Avan: Sebeos, 112.4-5; Hdip'sime: Sebeos, 121.5-28; St. Grigor at Dvin: Sebeos, 100.5-18.
134

56

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

decoration, and dedication of the "temple of the Lord" by Prince Juan'kr and test it against the evidence of the epigraphic record.'36 According to the narrative, Juangir stated that the church was built "as a refuge" for a fragment of "the living cross," a reference to the piece of the True Cross granted to him previously by Constans II.13' In depositing the fragment in the church that he had built, Juansir is imitating King Solomon, who constructed the Temple in Jerusalem as "the house of the Name of the Lord" and "a place for the ark, in which is the covenant of the Lord."'38 As suggested above, the prayer ofjuanser can only be interpreted through Solomon's own prayer of dedication (1 Kings 8). Juan'ir refers explicitly to the Temple of Solomon, inviting the Lord to fill his church with his Holy Spirit in the form of a cloud in the same way that the glory of the Lord had filled Solomon's Temple as a dark cloud.'39JuaniSr calls on the Lord to hear the supplications of "those bordering on the sea in the east" worshiping in his church and to grant their prayers.'40 Solomon prays in exactly the same way: "Hear the supplication of your servant and of your people Israel when they pray towards this place."'41 There can be no doubt that the two prayers are directly related to each other. Although the prayer ofJuanSer was inspired by that of Solomon, it departs from its biblical model in two important respects. In the first place, it is far more personal, articulating specific supplications for Juanser and his immediate family. By contrast, Solomon prayed that God would heed future requests from the people of Israel presented in the Temple or offered in the direction of the Temple when they were confronted by particular circumstances.'42 Solomon does not intercede for himself or his family. Second, Juanier'sprayer has an eschatological perspective. JuanSir asks for God's mercy "at the coming of the universal day of resurrection, when the invisible shall take the place of the visible, and the trumpet blasts shall summon the dead from sleep and the fiery/angelic hosts from the ends of the earth shall gather beside you."143 It is in the context of the Last Days that juansir intercedes for himself and "my spouse XosrovanuS, and my forefathers, especially my gray-haired father Varaz-Grigor and my mother Goriduxt and my brothers Varaz and Yezut Xosrov and my children."'44Whereas Solomon's prayer envisages certain Pero0 situations in the future, it does not contemplate the End of Time. The apocalyptic tone of Juangir's prayer finds contemporary corroboration in the History attributed to Sebeos.145 Although the passage expresses the past conduct and future expectations of a prince for himself and his family in much greater detail than is found in any inscription, there are several important correspondences between the literary and epigraphic evidence. As
Patmut'iwn Atuanic, 188.6-189.20; Dowsett, CaucasianAlbanians, 121-22. 182.4-5; Dowsett, CaucasianAlbanians, 116-17. 138 1 Kings 8:21. I am grateful to L. MacCoull for pointing out that Solomon's dedication prayer in 1 Kings 8 is read as a lection in most Eastern Christian liturgical rites for the dedication of a church. 139 Patmut'iwn Atuanic, 188.8-10: "send into this temple the grace of your holy spirit and after the manner of the temple of Solomon with the divine cloud of your glory like a fog, adorn and embellish this." Compare 1 Kings 8:10-13. 140 Patmut'iwn Atuanic, 188.10-11. This phrase implies a contemporary perception that the people of Aluank' had a sense of their own identity, defined in terms of both orientation and location.
137Patmut'iwn Atuanic,
141
142 143 136

1 Kings 8:30. 1 Kings 8:31-53.

Patmut'iwn Atuanic' 189.8-11; Dowsett, CaucasianAlbanians, 122. Atuanic4 189.12-15. 145 Greenwood,"Sebeos," 375-88.
144 Patmut'iwn

TIMOTHY GREENWOOD

57

noted above, the dedication ofjuanser's church is accorded a precise date, the eleventh day of the month of Arac'.146 juangir prays for his own safety in this world in the same way that Annay sought divine protection for herself and her children in the conclusion to the Bagaran inscription (A.3). He also prays for God's mercy on the Day ofJudgment for himself and his family. This matches the specific naming of close relatives found in the inscriptions of Bagaran (A.3), Alaman (A.4), Aru' (A.11), and T'alin (A.12), in which the founder names those members of his immediate family deserving of God's mercy and seeks intercession on their behalf. juanser's prayer concludes with an explicit statement as to why he had built the church: "Deliver me from eternal temptations and in return for this temple which I have built to your name on earth, prepare for this sinner a spiritual This stresses once again the personal benefit resting place in the house of your Father."'147 that was believed to accrue to the founder of a church. Indeed, the foundation is presented almost in contractual terms, whereby God will be required to grant Juanser heavenly citizenship in return for his earthly foundation. The precise coincidence of epigraphic and textual evidence shows that Juansir's prayer reflects the perceptions and sentiments of a contemporary. It follows that this prayer is of great importance when considering the motivation behind princely involvement in church foundations because it supplies additional information to that found in the inscriptions. Juansar prays for present protection and prosperity, mercy for his immediate family on Judgment Day, and membership of heaven for himself. There is no sense of ongoing benefit for future generations of Juanser's family. This is borne out in the epigraphic record and goes some way toward explaining why princely families were involved in multiple foundations. Aside from the prestige associated with such constructions, and the tradition that rival family members developed different centers, each prince had personal responsibility for his eternal fate and could not rely upon the achievements of his ancestors to secure heavenly citizenship on his behalf. There is one further intriguing element within Juanier's prayer. It seems to imply that the church was open to public worship: "Andif the whole country of Aluank' should come together to glorify your name in this house through the power of this cross that accepted Christ, may you close the gates of the enemies and impede the marauders."'48However, Solomon's prayer of dedication also contemplates the people of Israel praying and making supplication to the Lord in the Temple itself in times of defeat. In the light of this biblical parallel, the reference should not be treated as evidence for public participation in the services held in the church. The prayer ofjuangsr is followed by a short passage in which he promised on oath that he and his sons would lay "the first-fruits before the Lord and receive his mercy" at that place.'49 This action implies a personal, continuing attachment to this particular foundation on the part ofjuanger, one that is illustrated by his subsequent decoration of the church following his first visit to the King of the South, alias Mu'awiya.15o Frustratingly there are no parallel inscriptions recording subsequent endowments, either by the original founder or founders or their immediate successors.
146 147 148 149 150

Patmut'iwn Atuanic, 187.21; Dowsett, CaucasianAlbanians, 121. Patmut'iwn Atuanic' 189.18-20.

Ibid., 188.13-15. Ibid., 189.22-190.1.

Ibid., 192.8-196.10; Dowsett, CaucasianAlbanians, 124-26.

58

A CORPUS OF EARLYMEDIEVAL ARMENIAN INSCRIPTIONS

Before considering the other inscriptions in the corpus, let us examine these Armenian inscriptions from a different angle, namely, the insight they shed upon the anticipated function of the individual church. The churches at Tekor (A. 1), Hripcsime (A.2.1), and Nax'avan (A.8) are vkayarank'or martyria, housing relics of St. Sargis, St. Hrip'sime and St. Step'anos/Stephen the first martyr, respectively. The passage from the HistoryofAtuank' examined above reveals that the church founded by Juan'ir was designed to receive a fragment of the True Cross. Moreover, the small martyrium described in the same history as being built by the hermit Yovsip in the village of Ylerk in Arc'ax eventually contained relics of St. Step'anos, St. Georg, St. Andrew, the Apostle Thomas, and John the Baptist.'51 Thus it is clear that certain churches were perceived primarily in terms of their role as the repository of specified holy relics. The other Armenian inscriptions do not comment directly upon the function of the church. It may be significant, however, that the inscription at Aru' (A. 11) describes the building as surbkat'olikis,"this holy cathedral." The proximity between this church and a secular palace thirty yards to the south has already been noted. Their juxtaposition implies that this church had an important role in court ceremonial, a proposition supported by its impressive size, the number of windows (to maximize the amount of light entering the building), and the choice of subject for the principal wall painting in the eastern apse. Christ is portrayed standing on a pedestal and holding a parchment with a verse from John's Gospel.'52 The verse has an eschatological significance because it refers to a time when Christ will show himself to those who love him.153 It is possible that the onlooker is also being invited to interpret the presentation of Christ as king and judge in the context of the temporal authority enjoyed by Grigor Mamikonean over Armenia as the principal client of Mu'awiya. In other words, there is an intentional, albeit implicit, association between the subject of the wall painting and the temporal authority enjoyed by the sponsor of the church. Some churches, therefore, were conceived and functioned as palatine churches. By contrast, none of the inscriptions refers specifically to the role of a foundation as mausoleum. The literary sources, however, contain valuable information about Armenian princely burial practices, revealing that family members were usually buried in the same place. The Iranian origin of this tradition was first identified and discussed by N. G. Garsoian.154 The History attributed to Sebeos records that Smbat Bagratuni and his son Varaztiroc' were both buried at Dariwnk'.155 This tradition was still practiced at the end of the seventh century, when Lewond records that Alot Patrik Bagratuni was also buried there.156By way of later comparison, the History of Tcovma Arcruni records that several generations of Arcruni princes were buried in the monastery of the Holy Cross in the
Patmut'iwn Atuanic, 280.5-282.19; Dowsett, CaucasianAlbanians, 181-83. Der Nersessian, L'artarmenien,72 and figs. 47 and 48. 153 John 14:21: "Whoever has my commands and obeys them, he is the one who loves me. He who loves me will be loved by my Father and I too will love him and show myself to him." 154 N. G. Garsoian, "The Locus of the Death of Kings: Iranian Armenia-The Inverted Image," in TheArmenian Image in Historyand Literature,ed. R. G. Hovannisian, Studies in Near Eastern Culture and Society 3 (Malibu, Calif., 1981), repr. in N. G. Garsolan, Armeniabetween Byzantiumand the Sasanians (London, 1985), no. 43-44. xI, 155 Sebeos, 104.6-9 and 144.30-31. 156 Lewond, 18.20-24; Arzoumanian, Historyof Lewond,55.
152

151

TIMOTHY GREENWOOD

59

province of Albag, including the grandfather, father, mother, and elder brother of Gagik Arcruni who dominated Vaspurakan in the first half of the tenth century. 157The earliest extant inscription to specify that the original foundation was used as a mausoleum dates from the early eleventh century. It reveals that the monastery of Marmalen, founded in 1029 by Vahram prince of princes, anthypatos patrikios,and son of Grigor Pahlavuni, prince of Armenia, was "the place of our tomb."158In 1040 Apllarip, marzpanof Armenia, built the church of Grigor Pahlavuni/Apulamrenc' in Ani as a "sepulcher for my father Grigor and my brother Hamze and my sister Seda."'59 Clearly Armenian churches could fulfill a variety of functions, serving as martyria, as constituent elements within palatine complexes, and as sepulchers. However, although some foundation inscriptions mention the deposit of holy relics, they do not otherwise specify the purpose or purposes to which the church would be put. For the most part, the remaining inscriptions comprise simple invocations for divine assistance or forgiveness (A. 17 and A. 18), commemorations (B.3 and B.4), epitaphs (A. 13 and B.2), and graffiti (A.20).160 Five inscriptions, however, require further comment. The Greek inscription at Ereroyk' (A. 16) is a quotation of Psalm 93:5: "Holiness befits your house, O Lord, for evermore."161 None of the other inscriptions in Armenia cite the Bible directly; the closest is the allusion in the Mastara inscription (A. 10.3) to Christ as the bridegroom of the church. This direct citation coupled with the evident preference for Greek rather than Armenian places this church securely within a broader Near Eastern tradition. Ten similar inscriptions quoting this verse have been found scattered across the Near East from Thrace to Syria, Palestine, and the Negev; seven are in mosaic, two are carved on stone, and one is painted.162 One of the two stone inscriptions, from an unknown site in Cilicia, occurs on the upper face of a fragmentary chancel plaque. The other was incised on a dovetail plate in relief and located in the lintel of the eastern doorway in the southern
157 Agot Arcruni, grandfather of Gagik: T'ouma Arcruni, 217.19-23. Derenik Arcruni, father of Gagik: 228.16-19. Sop'i, mother of Gagik: 229.7-9. AUotArcruni, elder brother of Gagik: 249.5-10. It is also worth noting that the same monastery had previously received the remains of two great-uncles of Gagik: Grigor Arcruni (199.31-200.2) and Gurgen Arcruni (204.26-29). 158 Atiian, Sirak, 148. 159DHV 1: no. 97 and pl. xI. The word for sepulcher is 4,urnuiy,. 160 The inscription associated with the famous bird mosaic (B. 1) is also commemorative: "For the memory and salvation of all Armenians whose names the Lord knows." The precise meaning of the phrase "all Armenians" is unclear. Although theoretically it could be interpreted to mean all Armenians anywhere in space or time, on the balance of probabilities, it seems more likely that it was intended to have a much narrower meaning, defining Armenians within that particular community in Jerusalem. This implies that the community was mixed, comprising both Armenians and non-Armenians. 161 This numbering follows that of the Septuagint and the Armenian version. From Ps. 10:22 to Ps. 147.11, one has to subtract one to find the corresponding Psalm in the Hebrew and King James versions. 162 de Cilicie, ed. G. Dagron and D. Feissel, TM Monographies 4 For a complete bibliography, see Inscriptions (Paris, 1987), no. 23. All of the inscriptions cited are in mosaic with the exception of that in Thrace at Medeia and that being recorded by Dagron and Feissel under this entry. One further example, also in mosaic, has recently come to light, from Horvat Be'er Shema' (Birsame or Bersaman) in the southwest Negev: V. Tzaferis, "Ten Greek Inscriptions in Mosaic Pavements in the Ancient Christian Church of St. Stephen," Eretz-Israel 25 (1996): 81*-82* and fig. 9. The earliest dated mosaic inscription is that from Ascalon (A.D. 493): K. J. H. Vriezen, "Inscriptions in Mosaic Pavements in Byzantine Palestina/Arabia Quoting Texts from the Old Testament," in The Use of SacredBooksin theAncient World,ed. L. V. Rutgers and P W. van der Horst (Leuven, 1998), 251. The latest dated mosaic inscription of this series is that at Qabr Hiran (A.D. 575): E. Renan, Mission de PhInicie (Paris, 1864), 611.

60

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

wall of the North church at Deir Sem'an (Telanissos), close to the pilgrimage complex at Qal'at Sem'an in northern Syria.163 In terms of its content, layout, and position, this inscription is close to that found at Ereroyk'. Although a direct connection between these churches is unlikely, this epigraphic correspondence, noted by Shelov-Kovediaev, is significant.164 It corroborates the proposition that the basilica at Ereroyk' reflects Syrian traditions and practices in terms of church design and decoration.'65 It may also supply a very tentative terminus post quemfor the construction of Ereroyk', because development at Qal'at Sem'an, including the North church, began in earnest only after 476. As noted previously, the content of the three Arabic inscriptions at Zvart'noc' (A.19) remains at issue. Several readings proposed by Xac'atryan have been rejected by R. Hoyland, including the alleged references to the "collections of rents" and the "kharadjof the village."'166Nevertheless, those elements securely identified by Hoyland afford several important insights. The third inscription comprises an Arabic name coupled with the standard bismillah,inviting divine blessing upon the act; there can be no doubt that it was written by a Muslim Arab. Although the purpose behind the first inscription remains concealed in the first three lines of text, it contains an eighth-century hijra date, A.H. 154 (24 December 770/12 January 771) and correlates this with the governorship of Yazid b. Jarad and another figure. Such a synchronism within an Arabic inscription is unprecedented. The second inscription consists of an invocation by Ilyas b. 'Isa al-Waddaki, seeking divine pardon for himself, his whole family, his house, and his village. In addition to the possibility that his name may reflect a Christian Arab origin, the breadth of the intercession is also exceptional, extending to his whole family and even, it seems, to his village. As such, it bears an uncanny resemblance to intercessions expressed in Armenian inscriptions. Two inscriptions attest a secondary, subsequent bequest to an existing church or religious community. The first, unearthed recently in the Musrara quarter in Jerusalem (B.6), records that one Ewstat' eric, elder, was responsible for a mosaic floor and requests intercession for him and his brother Lukas. The second records the provision of a water source to T'alin by Uxtaytur (A.14). It is the first Armenian inscription to include a protective curse, warning that anyone who destroys the donation will become liable for the sins of the donor. In the wider context of Near Eastern epigraphy, this feature is not exceptionaltomb robbers are repeatedly anathematized, and an inscription at Aphrodisias threatens those who throw earth against a wall with condemnation by the 318 Fathers of the Council of Nicaea-but it seems to represent a new development in Armenia, reflecting a contemporary concern on the part of Uxtaytur for the survival of his bequest and, by implication, contemporary insecurity. 167 This leaves only the mysterious inscription of Ezr at Avan (A.6). The History attributed to Sebeos records the circumstances surrounding the foundation of that church in the Greek sector of Armenia after 591 by the so-called anti-katcolikos Yovhannes, described as

PPUAES IIIB: no. 1160; IGLSyr 2: no. 414. Shelov-Kovediaev, "Zametki,"63-65. cararmenienne,45-48; S. Ter-Gevorgyan, Hayastani ev Siriayivatk'ristoneakan 165Khatchatrian, L'architecture arncv 241-53. (Erevan, 2000), 242-50; Maranci, MedievalArmenianArchitecture, tarapetut'yan 'ut'yunnerW 166 XaC'atryan, Korpus,no. 6. 167 Dagron and Feissel, Inscriptions de Cilicie, no. 106.
163 164

TIMOTHY GREENWOOD

61

"one who was united in communion" with the Greek church over the status of the Council of Chalcedon.'68 At this time, the cathedral church of St. Grigor in Dvin had not yet been rebuilt following its destruction in 572.169 The vessels of that church had been taken into safekeeping in the city of Karin.'70 Thus neither the cathedral nor its vessels were available to Movses, the katcolikos in the Persian sector. The construction of the church at Avan should therefore be interpreted as part of a concerted attempt by the Greek church to gain control over its Armenian counterpart. These ambitions had to be set aside when the forces of Xosrov II drove the Romans out of Armenia in a series of hard-fought campaigns after 603, but they were not forgotten. In the aftermath of victory, Heraclius entered into negotiations with the kat'olikos Ezr on the issue of church union. The History attributed to Sebeos maintains that Ezr was threatened with a similar division of the see between the Greek and Persian sectors unless he entered into communion with the Greek church.171 Although Ezr is castigated in later Armenian histories for accepting that union, the relevant passages in this History are not hostile to him, suggesting that the author of these passages, like Ezr himself, belonged to that party within the Armenian church which preferred the Greek theological position on Chalcedon. This is the context in which Ezr's inscription at Avan needs to be interpreted. Arguably it is an assertion of his claim to possession of that church and the Christological tradition espoused by Yovhannes. The combination of Greek and Armenian words is unique, as is the unexpected use of cryptograms. It is significant that Ezr's successor as kat'olikos, Nerses III, also employed Greek in his inscriptions at Zvart'noc' and that he too was accused of pro-Greek sympathies. The choice of Greek was intentional, evoking association with and expressing sympathy for the Greek church. The Narratio de RebusArmeniae,completed in the year 690, attests the same Armenian tradition. G. Garitte observed: "Visiblement, notre auteur n'a pas l'intention d'&crirepour des lecteurs non armeniens; nulle part il ne manifeste le moindre souci de se faire comprendre d'un public byzantin; en de nombreux passages, il aurait du, si tel avait 6te son but, s'exprimer autrement ou fournir des renseignements qu'il ne pouvait supposer connus hors des frontifres de l'Armenie."172 Having proposed that the anonymous author was both Chalcedonian and Armenian, Garitte nevertheless envisaged that the original was composed in Armenian and translated into Greek, noting that "les fautes de traduction et les armenismes qui apparaissent dans son texte peuvent s'expliquer aussi bien chez un traducteur grec trop attach6 a son modble arm6nien que chez un Armenien peu sensible a certaines nuances de la langue grecque."'7"He preferred the second hypothesis. However, the study of the inscriptions suggests a third possibility, that the text was composed very deliberately in Greek by an Armenian and that there was no intermediate, translation stage in the process of transmission.

Sebeos, 91.11-24. For its destruction, see Sebeos, 68.5-12; for its rebuilding, at the instigation and insistence of Smbat and the marzpan,see Sebeos, Bagratuni, and against the wishes of the commander of the fortress (berdakalin) 100.5-18. 170 Sebeos, 91.21-23. 17' Ibid., 131.31-132.11.
169 172

168

173

361. Garitte,Narratio, Ibid., 400.

62

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

HISTORICAL POTENTIAL

Since these inscriptions are original and manifestly have not been amended or altered, they constitute important historical sources. In the first place, they represent an invaluable record of contemporary social terminology. As such, they may be used as "controls" against which to test the terminology found in the historical sources which date from this period. This comparative exercise indicates that the latter have not been subjected to later interference or rewriting in the course of transmission-a particular concern in the light of the late manuscript tradition. At the same time, these inscriptions comprise independent freestanding historical sources in their own right and therefore provide key information about the structure and functioning of Armenian society, at least at an dlite level, under Sasanian, Byzantine, and Islamic rule. Finally, they attest something of the impact upon Armenia of the dramatic political upheaval experienced in the Near East during the seventh century and the reactions of Armenians to those events. When the range of Persian, Greek, and Armenian honors and titles preserved in the epigraphic record is compared with that found in the History attributed to Sebeos, the History of Atuank, and other works, it is found to correspond closely. This coincidence supports the contention that these historical compilations preserve contemporary materials in an unaltered state and therefore reflect seventh-century conditions and perceptions. Thus the inscription at Bagaran (A.3) asserts that the church was completed three years after the thirty-eighth year of Xosrov, that is, June 629/630, at a time when Varaztiroccaspet was marzpanor governor ofArmenia. The History attributed to Sebeos records that Varaztiroc' was appointed marzpanofArmenia by Xosrov's short-lived successor, Kawat II, in 628, and that he fled from Armenia following "a great altercation" with Rostom, prince of Atrpatakan, in the winter of either 631-632 or, more probably, 632-633.174 The inscription confirms that he held the office reported in the text at the relevant time. It also reveals that the hereditary Bagratuni title aspet was still current. This corroborates the evidence in the History attributed to Sebeos, where Varaztiroc' is regularly identified simply as the aspet without being named."75 The epigraphic record is no less important in recording the award of Byzantine honorific titles to Armenians.'76 The inscription at Mren (A.7) opens with a reference to the alland sparapetof Armenia and Asorik'-a unique combipraiseworthy patrikios,kouropalates, nation of honors which could belong only to the preeminent figure within Armenia at this time. By contrast, the founder of the church at Alaman (A.4) is identified as Grigor etustr, the Armenian form of the middle-ranking imperial title of illustris.'77Finally, the tomb174Appointment: Sebeos, 128.36-129.2. Flight: Sebeos, 132.12-35 and Thomson and Howard-Johnston, Sebeos,2:229. 175Greenwood, "Sebeos," 355-56 and n. 125. 176 In the interests of clarity, I have adopted the conventions followed in ODB for Byzantine titles and offices rather than the Armenian transliterations of them as found in the inscriptions. The latter are inconsistent. However, for the sake of precision, the translations supplied in Appendix 1 do contain the exact transliterations. 177 For a discussion of the titles awarded to members of the senatorial aristocracy and changes over time, see J. E Haldon, Byzantiumin the SeventhCentury,rev. ed. (Cambridge, 1997), 161-72. Haldon maintains that the senatorial rank ofillustris came to be supplanted during the reign ofJustinian I by new, higher grades, ofmagand that the emperor awarded the title of illustris to a number of provincial nificus, gloriosus, and gloriosissimus officials and even curialeswho had previously borne only the ranks ofspectabilisand clarissimus.This downward shift in the status of illustris may also be observed within the Miracula Sancti Demetrii,which identifies a man of

TIMOTHY GREENWOOD

63

stone of the blessed tar/lordArtavazd Kamsarakan (A.13) recorded that he held the titles apo hypaton,patrikios, and prince of Armenia/i~xan Hayoc, and that his father Hrahat had been patrikiosand terSirakayew Asaruneac'/lordof Sirak and AUarunik'.These inscriptions reveal that Armenians were awarded a range of Byzantine imperial titles, implying familiarity with a hierarchy of honors.'78 The historical texts attest a similar variety. The History attributed to Sebeos records that three prominent Armenians were accorded the title of kouropalates-Dawit' Sahaiuni, Varaztiroc', and Hamazasp Mamikonean-and that several The passage recording the elevation of Hamazasp ter others were appointed patrikios. 179 to the high rank of kouropalates of the Mamikoneans Mamikoneic'/lord by Constans II in 655 also reports that the emperor granted "honors" to other princes, including, it seems, imperial titles.'80 The History of Atuank'supports this interpretation. It records that on submitting to Constans II, the prince of Aluank', juangir, received from him the titles of apo as well as the unique arewelicc hypaton,protopatrikios, kusakal/governor of the East.'81Moreover, the passage alleges that juansir was accorded a remarkable discretion by Constans II, to distribute imperial titles as he thought appropriate: "Andhe made him protopatrikios and had conveyed to him the honor of patrikiosand hypatosand apo hypatonand stratelates and elist/illustrisfor one thousand two hundred men, so that to whom he himself might It seems, therefore, that Constans II entrusted his princiwish, he could bestow that."'182 in the eastern client Caucasus, juangsr, with the responsibility for attracting putative pal imperial clients through the concession of a range of titles, from patrikiosto illustris. The coincidence of etustr in the Alaman inscription (A.4) and elist in the above list is striking, confirming the distribution of lesser imperial titles. As will be discussed below, there seems to have been a concerted effort on the part of both Heraclius and Constans II to attract a range of supporters from the Armenian, Albanian, and Persian elite and to tie them into the ramified political structures of the empire through the grant of imperial honors. However, the inscriptions are not solely concerned with foreign titles and honors. They also corroborate the contemporary use of three Armenian social terms: tar,translated "lord of the family,"and isxan as "lord," "prince of Armenia." All three featanutr,, Hayoc; The ture in the attributed to first textual reference to a prince with a Sebeos.'83 History wider political authority over Armenia occurs in connection with Dawit' Sahaiuni; he is
good birth as one of 'the so-called illustrioi':P Lemerle, Lesplus anciens recueilsdes Miraclesde saint Demitrius, 2 vols. (Paris, 1979), 1:161. See "Hypatos," ODB 2:963; "Illustris," 2:986; "Kouropalates," 2:1157; and "Patrikios,"3:1600. 178 179 Appointments as kouropalates:Sebeos, 133.32 (Dawit' Sahairuni); 144.25-29 (Varaztiroc'); and 175.10 (Hamazasp Mamikonean). Appointments as patrikios: Sebeos, 67.23 (Vard Mamikonean); 104.23 (Atat Xorxoiuni); 132.31-32 (Varaztiroc'); and 139.5-6 (T'eodoros RPtuni). 180 Sebeos, 175.9-12. 181 Patmutziwn Atuanic4 182.9-10; Dowsett, CaucasianAlbanians, 117. The title pioton patrik, that is protopatrikios,implies a singular precedence. The History attributed to Sebeos records that in ca. 632 Heraclius honored Varaztiroc' "more than all the patriks who were in his kingdom," a phrase which implies that he received a similar honor: Sebeos, 132.31-32. The history of this title is obscure, beyond the fact that it emerged in the and comesof Opsikion was the last holder of the title (killed late 4th century and that Varazbakurprotopatrikios in 711); see Theophanes, Chronographia,1:380.14-15 and A. B. Karamaloude, "Haparflp1 too6 ciF; G' Oj In 5 161-68. the 7th the award of this t6 (1983): century, singular Cpo )ooxatpucioJ p6tptgo tthudvztvo," ,Vlteltwcra honor seems to be restricted to prominent client princes from the Caucasus. In this respect, it mirrors the award of the title kouropalates.I am very grateful to L. MacCoull for pointing out that this may be compared with the 3d century Palmyrene Vaballathus, named Corrector totius Orientis. 182Patmut'iwn Atuanic, 182.2-4; Dowsett, CaucasianAlbanians, 116-17. 183 For the frequency of tir and tanutir, see the entries for "lord" and "tanutir/headship" in Thomson and

64

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

appointed as izxan i verayamenaynasxarhac'n,"prince over all the countries," as well as being given the title of kouropalates.'84 Although it is incomplete, the inscription at Mren (A.7) seems to confirm the award of both titles of Dawitc.The coincidence in social terminology between the literary and epigraphic sources even extends to the parallel omission of one notable term, namely, This is conspicuous by its absence from both the latest nonaxarar. tices in the History attributed to Sebeos and the corpus of Armenian inscriptions.'85 The exact meaning of these terms and the significance of their frequency will be discussed below. For the present it is sufficient to note the close coincidence in terminology across the two classes of primary material, supporting the proposition that these Armenian historical texts have retained their original form in spite of their late manuscript traditions. This initial survey prompts one further observation. Both the inscriptions and the Histories reveal that an individual Armenian prince could be defined in several ways at the same time. Thus Hamazasp Mamikonean is described in the History attributed to Sebeos as lord of the Mamikoneans, kouropalates, and someone holding the office of prince of the of he is presented in terms of precedence within the Armenia.'86 In other words, country Mamikonean house, in terms of his Byzantine title, and in terms of his political headship of Armenia as client of Constans II. The choice of which title to use within an inscription is therefore very significant, reflecting how contemporaries perceived themselves and their identity or wished others to recognize their status. The inscription at Alaman (A.4) is particularly revealing in this respect; the founder Grigor chose very deliberately to employ his imperial honor in preference to an Armenian social term. By contrast, Nerseh is defined in the synchronism simply as tdrSirakayew Asvaruneac, "lord of Sirak and Asarureference A without to similar any Byzantine title.'87 preference may also be detected nik'," when imperial titles occur in conjunction with Armenian terms. At Tcalin (A. 12), Nerseh is defined as apo hypaton and patrikios before he is identified as ter Sirakay ew Asaruneac, whereas at Nax'avan/Noramanuk (A. 13), Hrahat is titled patrikios ahead of tir Sirakayew Asaruneac. This order may reflect the relative significance attached to each element. Three of the inscriptions contribute directly to the study of Armenian social structure through their use of complex synchronisms. Far from being a random collection of indicators, the sequence of components within each synchronism was carefully devised to reHoward-Johnston, Sebeos,2:324 and 330. For ihxanHayoc4see Sebeos, 138.33, 143.24-25, 169.14-15 (all relating to T'eodoros Rituni) and 174.29 and 175.11 (both referring to Hamazasp Mamikonean).
184 185

Sebeos, 133.31-32.

The last reference appears in Sebeos, 149.7. 186 Ibid., 175.10-11. 187 The Long Recension of the geographical text attributed to Ananias of Sirak, the refers to Ahxarhacbyc' one imastunnNersehKamsarakan ew "the wise Nerseh Avoruneac'tirn, Kamsarakan, patrikn Sirakay patrik, lord of Sirak and Alorunik'": Ahxarhac'oyc'Movsesi Xorenac'woy. Gtographiede Moise de Corine, ed. and trans. A. Soukry (Venice, 1881), facsimile reproduction, with new introduction by R. H. Hewsen (Delmar, N.Y., 1994), 18.2829; R. H. Hewsen, The Geography of Ananias of Sirak(Afxarhac'yc').TheLong and ShortRecensions,Beihefte zum Atlas Vorderen des Reihe B 77 (Wiesbaden, 1991), 50 and n. 31. Although Hewsen equates Orients, Tiubinger this Nerseh with the Nerseh Kamsarakan who was appointed isxan Hayoc'byJustinian II at the end of the 7th century, there appears to be no reason why he should not be identified with that Nerseh terSirakayewAsaruneac named in the inscriptions at Alaman (A.4) and Mren (A.7). This complements Hewsen's statement that "L, the original recension of the ASH, was written within a forty-five-year period, between 591 and 636" (Hewsen, 33). However, it should also be noted that the reference occurs at the end of a passage describing Geography, the cities of Africa. This context is difficult to explain, prompting Hewsen to conclude that it is a later interpolation (ibid., 96 n. 31).

TIMOTHY GREENWOOD

65

flect the existing social order in the vicinity of the foundation as perceived by a contemporary. The synchronism at Bagaran (A.3) is arranged as follows: Persian king, marzpan, local Armenian lord (ter),and finally a cleric, although the damaged condition of this part of the inscription makes it unclear whether this figure was the kat'olikos Ezr, as Orbeli maintained, or the local bishop.188Thus there is a hierarchy, stretching from a supranational figure to his governor and then the Armenian lord in control of the district. The inscriptions at Alaman (A.4) and Mren (A.7) broadly follow this sequence, although both reveal intriguing variation. At Alaman, the sequence is emperor, local Armenian lord, and then local bishop. Significantly, the inscription omits any intermediary figure between the emperor, Heraclius, and the local lord, Nerseh, implying a direct relationship between them. The Mren synchronism, dating from two years later, is similar to that at Bagaran in structure, albeit with a Byzantine rather than Persian orientation: emperor, the authority ew Asorwoc, of someone who was simultaneously patrikios,kouropalates,and sparapetHayoc" local bishop, and local tanutir. Once again, the social order is presented as operating on three levels: emperor, principal imperial representative, and local lord and bishop. On the basis of this evidence, it seems that after 638 Heraclius sought to control Armenia through a single, designated client. This approach to the government of Armenia was maintained during the first decade of the reign of Constans II. Only when TVeodoros Rituni transferred his allegiance to Mu'awiya in 652 did individual princes once again seek direct contact with the emperor. The History attributed to Sebeos records how Constans II advanced to Karin the following year and there received the submission of a disparate group of Armenian princes and their armed forces, comprising those who were from Fourth Armenia and those who had left "the kRtunian"-in other words those who had repudiated the authority ofT'eodoros Rituni following his political volte-face and wished to renew their ties with Constans II.189 This state of affairs, whereby Armenian princes entered into a direct relationship with the emperor, was short-lived, lasting only until his appointment of Musel That new office Mamikonean as isxan Hayoc'hecelocin, "prince of the Armenian cavalry."'190 challenged the status of T'eodoros RItuni as isxan Hayoccand seems to reflect a general power of command over those Armenian princes who had refused to collaborate with Mucawiya. Intriguingly, this innovation did not outlive T'eodoros. After 654, the sources report the intense competition between the outside powers to appoint their own client to the office ofisxan Hayoc'.Unfortunately it is at this moment that contacts and relationships at a local level disappear from view. With this framework in mind, let us now turn to examine the social status of the individual sponsors and the honorific titles associated with them. The church at Bagaran (A.3) was founded by the local lord, But Airuelean. He is identified solely in terms of the Armenian social term tr. This term designates the head of an Armenian noble family within a district or districts under the control of that house. It is always found in an Armenian
188 Orbeli, Izbrannye trudy, 376-77. Orbeli proposed that this part of the inscription had originally named the kat'olikos Ezr but that this reference had been deliberately erased by someone opposed to his pro-Greek, Chalcedonian views. However, the structure of the other Armenian inscriptions, all of which identify the local bishop in this context, undermines this contention. The damage need not be intentional but may simply be caused by weathering. 189 Sebeos, 165.17-31. 190 Ibid., 166.17-18.

66

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

context and is always applied to an Armenian. It was never created or conferred by an external power. Although such powers might seek to influence who held the title, they did not determine whether or not someone could lay claim to that title. That right was based upon kinship. The inscription at Alaman (A.4) records that the small church was founded by Grigor etustr and his wife, Mariam, in October 636/637. As the synchronism refers specifically to Nerseh as ter Sirakayew Asaruneac, it is clear that Grigor did not hold the status of lord of the district in which his foundation was located. At the same time, however, the very act of founding a church attests a strong attachment between Grigor and the site. Frustratingly, the inscription does not indicate the basis for that attachment, but it seems most likely that Grigor held responsibility within the district under the lordship of Nerseh. Although his family origin and Armenian social status remain unknown, as noted previously, Grigor held the honorific rank illustris. It seems, therefore, that Grigor preferred to define his status in terms of this middle-ranking imperial title, as someone who belonged within the extended political order of the Byzantine Empire. This inscription is significant for three principal reasons. It indicates that someone other than the local lord nevertheless had sufficient capital resources to found a church, albeit of extremely modest size. It confirms that imperial titles were distributed beneath the level of the local lord, attesting a considerable down-reach. Third, as noted above, when the church was completed in 636/637, there was no single figure who was recognized in this district of Armenia acting as an intermediary between emperor and local lord."19 The identity of the founder of Mren remains contentious. Although Orbeli and others were correct in identifying the anonymous patrikios, kouropalates,and sparapetHayoccew Asorwoc' as DawitcSahaiuni, it is worth remembering that this reference forms part of the synchronism rather than the dedication and thus has no obvious role in determining the identity of the sponsor. 192 Indeed, the scope of the intercession appears to run counter to the proposition that Dawit' was responsible. The church is described as being for the benefit of the "Kamsarakank', Mren and all [?],"suggesting that a member of this house was responsible for its construction. There is, however, no extant inscription in which the original benefactor is also mentioned in the synchronism, and so it is unclear how such duplication would have been tackled. Intriguingly, the synchronism at Mren applies a very positive epithet to the unnamed individual; in addition to listing the various honors and offices he held, the inscription also describes him as amenagov,"all-praiseworthy."The fact that such an epithet qualifies an Armenian figure rather than an emperor is unique and may reflect, albeit implicitly, the direct involvement of Dawitc Saharuni in the foundation. Moreover, it is conceivable that Dawit' Sahalruniwas himself ajunior member of the Kamsarakan house, promoted by Heraclius to serve as his principal client within Armenia. This
This suggests that the authority of M4ie Gnuni, the zawravaror general of Armenia at this time, was not recognized in every district of Armenia. It is worth recalling that Sasanian influence persisted in Armenia much later than one might have expected-the sparapet,MuSel Mamikonean, and Grigor, lord of Siwnik', answered the call to arms in autumn 637 and died in battle at Qddisiyya on 6 January 638: Patmutiwn Atuanic, 173.16-20; Dowsett, CaucasianAlbanians, 110; and Sebeos, 137.13-20. Armenian ties with Persia were severed at the end of the 630s and not the beginning of this decade. This is the context in which Dawit' Saharuni was appointed "prince over all the countries of Armenia": Sebeos, 133.31-32. 192 Dawit' Saharuni is titled and identified as responsible for this church in the History of John kouropalates Catholicos: Yovhannes Drasxanakertc'i, 80.20-26; Maksoudian, Yovhannes 19.2. Drasxanakertc'i,
191

TIMOTHY GREENWOOD

67

would justify the range of new titles accorded to him as well as explaining the circumstances in which the existing head of the Kamsarakan house, Nerseh, could be named in the synchronism without apparently having a role in the foundation. It is worth noting that while Nerseh is defined as tr Sirakayew Asaruneac, he is also accorded the title tanuti. Tanutiris much less frequently employed than tir but seems to have had a broadly similar meaning, identifying the head of a single house. It is clear from this inscription that Nerseh was both tanuterand tir at the same time. The difference between the two terms is hard to explain, but it may revolve around the issue of external recognition. In the History attributed to Sebeos, tanutir occurs only in circumstances where the headship of a house is being confirmed by an external authority-usually Xosrov II, although on the last occasion it is Constans II. Thus, in my view, tanut&r identifies a tir whose headship has been specifically recognized by an outside power. The dedication inscription at Aru' (A.11) is not beset with such uncertainties. It asserts confidently that the church was founded by Grigor Mamikonean, isxan Hayoc, and his wife, Heline. Although Grigor is identified as a Mamikonean, his exact status within that house is not specified. Instead he is described simply as "prince of Armenia." This title expresses a much broader political authority within Armenia, one that was not restricted to a single family or a particular district. It was used to denote the principal client of an external power through whom Armenia was governed and Armenian military potential exploited. In this instance it reflects Grigor's role as the principal Armenian client of Mu'awiya. The inscription appears to reflect a desire on the part of Grigor to be defined in terms of his Mamikonean origin and in terms of his premier Armenian rank. As such, it may be compared usefully with the later inscriptions from T'alin (A.12) and Naxvavan/ Noramanuk (A.13). In the former, Nerseh is described in terms of his Byzantine titles and as tir Sirakayew Asaruneac';by contrast the latter defines his grandson Artavazd Kamsarakan in terms of his imperial honors once again and as isxan Hayocc. When these three are examined a can be detected. The two inscriptions collectively, pattern figures holding the title "prince of Armenia" are both defined in terms of their family origin-Mamikonean and Kamsarakan respectively-but are not accorded specific status within these houses. By contrast the two figures who did not hold the title isxan Hayoc, namely, Nerseh and his son Hrahat, are both described as tir SirakayewAsaruneac' but are not identified explicitly as members of the Kamsarakan family. Evidently the convention was to use the foremost title, and isxan Hayoc'took priority over It does not follow that Grigor and Artir. tavazd did not also enjoy headship of their houses, but rather that tir was the less significant title in the context of Armenian social terminology and therefore omitted. Indeed, the apparent determination of Grigor and Artavazd to specify their Mamikonean and Kamsarakan origins seems to reflect how keenly these identities were felt. The inscriptions ofT'alin (A. 12) and Naxiavan/Noramanuk (A. 13) yield valuable findings of their own because they link together three generations of Kamsarakank': Nerseh, Hrahat, and Artavazd. Both their imperial and Armenian titles repay close investigation. As we have noted, Nerseh was titled apo hypaton,patrikios, and tr Sirakayew Asaruneac'; Hrahat was merely patrikiosand tr Sirakay ewAlaruneac'; while the blessed tr Artavazd Kamsarakan held the titles apo hypatonpatrikiosand i~xanHayocc.Following standard Byzantine awards that did not pass automatically practice, imperial honors were personal, ad hominem from father to son but had to be conferred individually. It is significant that Artavazd

68

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

regained the rank of apo hypaton,previously enjoyed by his grandfather Nerseh but not by his father Hrahat. These very precise changes in entitlement point to continued Byzantine influence within those northwestern districts of Armenia under Kamsarakan control during the first half of the eighth century. The inscription at Nax'avan/Noramanuk (A. 13) is also significant because it represents the only extant record of the elevation of Artavazd to the rank of isxan Hayoc. The period of his tenure remains conjectural but should probably be placed between 720 and 760.'93 This inscription suggests that the record of appointments to this office in the eighth century is incomplete. However, there is an alternative solution. The combination of this rank with two imperial honors implies that Artavazd may have been promoted to this office by a Byzantine emperor, as a rival to the caliph's preferred client. Far from being unprecedented, the elevation of a rival figure to the same office was familiar practice in the constant competition for advantage within Armenia played out by opposing powers. This would explain why Artavazd failed to register in the standard sequence of isxank'Hayoc'.194 Hitherto this study of Armenian social structure has been focused primarily upon the male 6lite. However, the Armenian inscriptions also reveal something of the powers and responsibilities of noble-born women in early medieval Armenia, as wives, widows, and figures of importance in their own right. Six of the Armenian inscriptions from Armenia identify the wife of the founder: the unnamed wife of Sahak Kamsarakan at Tekor (A. 1); Annay, widow of But Aiuelean, at Bagaran (A.3); Mariam, wife of Grigor etustr,at Alaman (A.4); Heline, wife of Grigor Mamikonean, at Aruc' (A. 11); Su'an, wife of Nerseh, at T'alin (A. 12); and Su'an Mamikonean, wife of Artavazd Kamsarakan, at Naxcavan/Noramanuk (A. 13). It could be argued that these inscriptions simply reflect the proper concern of a Christian husband for the spiritual welfare of his wife, as expressed in Ephesians 5:25, and thus they should not be interpreted as proving joint responsibility for these constructions.'95 However, the inscription at Alaman (A.4) employs a first person plural sinec'ak, "we built," supporting the contention that wives could play a more active role than has previously been supposed. The inscription at Bagaran (A.3) confirms this. After the murder of But, the text reveals that his widow, Annay, took over responsibility for the completion of the church, a task that took three years. There can be little doubt that Annay retained control of sufficient material and human resources to enable the construction to be completed. The capacity in which Annay controlled these resources is unclear. Was it as widow, as guardian for her children, or as a noble-born daughter with lands and resources of her own? The additional intercession appended by Annay to the original conclusion is also instructive. In asking for God's protection over herself and her children, Annay seems to be indicating her present concern that they were all in danger. Unfortunately the outcome of these events is-and will remain-obscure. The epitaph to Artavazd Kamsarakan at Naxiavan/Noramanuk (A.13) also contains invaluable information. Artavazd is described as the husband of Sulan Mamikonean. This
For a full discussion of the date of this inscription, see below. See A. N. Ter-Ghevondian, "Le prince d'Armenie a l'6poque de la domination arabe," REArm3 (1966): 185-200; J. Laurent, L'ArminieentreByzanceet l'Islamdepuis la conquitearabejusqu'en886, Nouvelle 6dition revue et mise A jour par M. Canard (Lisbon, 1980), app. 4.1, "Le Prince d'Arm6nie ou chef 'Arm6nien' de 1'Arm6nie du VIIe au IXe siecle," 400-407. 195 Eph. 5:25: "Husbands, love your wives just as Christ loved the church and gave himself up for her."
193 194

TIMOTHY GREENWOOD

69

statement is significant in two respects. Instead of Su'an being described as the wife of Artavazd, Artavazd is described as the husband of Su'an. This suggests that the epitaph was commissioned by Su'an, who had outlived her husband. Second, Su'an is defined in terms of her family origin, as a Mamikonean, rather than in terms of her kinship by marriage, as a Kamsarakan. Again this reassertion of her earlier, Mamikonean, identity strongly implies that Susan was alive when the inscription was carved and that she was involved in its composition. It is unclear why she should decide to revert back to her original Mamikonean kinship, but it is possible that she expected or hoped to benefit from the ongoing protection of that house. Admittedly these inscriptions supply very limited information about the responsibilities of noble-born women in Armenia in the period. However, they provide a context within which three discrete narratives may be interpreted. Book 2, chapter 38 of the History of Atuank'records the intervention of Heline, princess of Great Armenia and wife of Grigor Mamikonean, in securing relics of Grigor the Illuminator for Bishop Israydl. He had been sent on a diplomatic mission by Varaz Trdat, prince of Aluank', to the kat'olikos of Armenia, Sahak, and the "pious prince of the Ayraratean district," Grigor.196 Helino interceded successfully with the kat'olikos and her husband on behalf of Israyel, allegedly because she was from the nahang or province of Aluank'. This account therefore depicts Heline taking an active role in church politics. The same History later records how queen Spram of Aluank' also became embroiled in theological controversy in the first decade of the eighth century.'97 While her husband Varaz Trdat was detained in Constantinople, Spram is presented as one of the two key figures in the political structure of Aluank'. Accusations of heretical belief were brought against her and provided the opportunity for Seroy patrikios and isxan Atuanic, prince of Aluank', to mount a successful coup. Such charges enabled Seroy to engineer the removal of both Spram and the katcolikos of Aluank', Nersos, at the same time, leaving the way clear for him to seize power. Spram is therefore depicted as acting on behalf of her husband during his enforced absence, in much the same way that Annay took over responsibility for the construction of the church at Bagaran after her husband's murder. The final passage, from the History of Lewond, records the perplexing conduct of Princess Sulan following the Armenian victory at Vardanakert in 703.198 When faced with a band of three hundred Arab soldiers being pursued by Armenian forces, she resolved to offer them sanctuary until they had recovered sufficiently, at which point she gave them packhorses from her own herd for onward transit to the caliph, 'Abd al-Malik. Again, therefore, the narrative presents a noblewoman playing a significant political role, adopting an independent stance, negotiating directly with the caliph and being rewarded by him. Evidently her decision was respected by the pursuing Armenian forces; they did not violate the sanctuary provided by her. The account also contains the fascinating aside that
Patmut'iwn Atuanic, 235.22-237.8; Dowsett, CaucasianAlbanians, 152-53. Patmut(iwn Atuanic, 293.4-311.5.20; Dowsett, CaucasianAlbanians, 189-202. These chapters chart the circumstances surrounding the deposition of Nerses and the terms of the settlement that was eventually achieved. Although there must be some doubt as to the authenticity of the correspondence between the kat'olikos of Armenia, Elia, and the caliph, 'Abd al-Malik, chapters 8-11 comprise the set of interlocking documents recording the settlement of the affair and form an exhaustive testamentary witness, shedding vital insight on contemporary social structure, land transfer inter vivos by both sale and grant, and dispute resolution. 198 Lewond, 25.19-26.3; Arzoumanian, Historyof Lewond, 60.
197 196

70

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

she owned a herd of packhorses. Such an incidental detail is rarely provided in any account, let alone one that is focused upon the actions and decision of a princess. Even if one harbors reservations about the accuracy of the episode, it proves that the author of the passage had no hesitation in ascribing such actions to a princess, implying that they could operate as independent political figures. Again, the context in which Princess Susan exercised leadership is missing; was she acting as widow, as custodian for her children, or as the sole heir to a district or region? Her princely origin is also unspecified. Although she has been identified as the wife of that Nerseh who founded the church at T'alin (A.12), this is clearly mistaken. The relevant colophon, quoted in full above, reports that Su'an Kamsarakan was martyred in Xaian/IHarran in Mesopotamia in 706/707, and it seems highly likely that it is this Su'an who was the wife of Nerseh.'99 It is hard to envisage the circumstances in which the same person could have saved the lives of defeated Arab troops and been rewarded accordingly by the caliph in 703 and yet endured a martyr's death some three years later. Thus it seems easiest to interpret these two passages as referring to different individuals, both named Susan. Thus the inscriptions shed fresh insight into the leadership responsibilities undertaken by several noblewomen. They confirm the impression that emerges from the literary sources, that women could and did play a significant role in Armenian political life, controlling resources and making independent decisions. These activities are largely hidden from view and rarely register in the primary sources.
ARMENIA AND THE GREAT POWERS: SOCIAL EVOLUTION

Having examined the Armenian inscriptions as contemporary records of social terminology, let us now trace the impact of Sasanian, Byzantine, and Islamic rule upon Armenia as reflected through the whole corpus of inscriptions. The inscriptions at Tekor (A. 1) and Bagaran (A.3) illustrate two features of Sasanian provincial administration as it operated within Armenia before 628. The first, earlier part of the Tekor inscription, dated on the basis of the reference to the katcolikos Yohan (ca. 478-490), notes the involvement of Manan hazarapetin the foundation. H. Aiairean maintained that Manan should be identified as Veh-Vehnam, named by Lazar as hazarapetofArmenia in 482 and 484.200 Whether or not one accepts his contention, it is clear from that text and the Buzandaran that the hazarapetacted in concert with the marzpanand that he had administrative responsibilities.201 This office was usually held by Persians. In times of internal crisis, however, it could revert to an Armenian, as happened in 485 when Vahan Mamikonean was invested by Valarq following the death of Peroz in battle against the
199 Yovsep'yan, Yidatakarankno. 17. H. AMairean, Hayoc'Anjnanunneri Baiaran, 5 vols. (Erevan, 1942-62; repr. Beirut, 1972), 3:193 and 5:119; Lazar P'arpec'i, Patmut'iwn Hayoc'ewTi'ut'ai VahanMamikonean,ed. G. Ter-MkrtcEeanand S. Malxasean (Tiflis, 1904; repr. Delmar, N.Y., 1985), 118.39 and 164.20-21; trans. and commentary by R. W. Thomson, The Historyof Lazar P'arpec'i,Occasional Papers and Proceedings 4 (Atlanta, 1991), ?66 and ?91. 201 Thus in Lazar's History, Veh-vehnam hazarapetappears only in the context of his support for other Persian officials, namely the marzpanof Armenia, Atrvinasp Yozmandean (in 482) and then Nixor Vinaspdat (in Patmut(iwn 484). Buzandaran (attributed to P'awstos), P'awstosi Hayoc4ed. K'. Patkanean (St. PetersBuzandac'woy with introduction facsimile 1883; Tiflis, 1912), by N. G. Garsoian (Delmar, N.Y., reproduction, repr. burg, 1984), 26.3-5 and 56.25-28; trans. and extensive commentary by N. G. Garsoian, The Epic Histories(BuzanHarvard Armenian Texts and Studies 8 (Cambridge, Mass., 1989), III.xii, Iv.ii, and 531-32. daranpatmut'iwnk'),
200

TIMOTHY GREENWOOD

71

The record of Manan's involvement in the construction Hephthalites the previous year.202 at Tekor is therefore intriguing. The inscription at Bagaran (A.3) also attests Sasanian influence, albeit in a different way. As we have seen, it employs regnal years of Xosrov II to locate the construction in time. Underlying this chronological choice is an implicit recognition of the sovereignty of the Persian king, that time itself was calculated by reference to his reign. However, the Bagaran inscription also provides useful evidence of the contemporary stresses within the Sasanian Empire after the death of Xosrov II. His successor, Kawat II, appointed a prominent Armenian, Varaztiroc', as marzpan.This broke with the convention, noted above, that Armenians should not hold the highest administrative office in their own country. Moreover, the decision to use the murder of Lord But Aruelean, securely dated to the thirty-eighth year of Xosrov II, as the starting point from which to reach forwards in time, is unprecedented. It implies that this was a period of chronological uncertainty within Armenia, at the very time when there was a rapid succession of short-lived kings and queens of Persia. After 628, Armenia was partitioned, most probably on the basis of the frontier agreed in 387.203 However, it seems that the boundary was twice redrawn in favor of Heraclius, once as part of his pact with Xoream/Sahrvaraz in July 629, and again following Xoream's assassination in June 630, when the new regime under Boran was desperate to secure peace.204J. D. Howard-Johnston has argued persuasively that the agreement negotiated by Boran's envoys at Beroea restored the frontier imposed by Maurice on Xosrov II in 591.205This state of affairs is corroborated by Ezr's request to the emperor for the salt mines of Kolb in 631.206 Although we cannot tell whether the kat'olikos was seeking absolute possession or an interest in the income or tax receipts, it is clear that such a concession would not have been within the grant of Heraclius unless the frontier had been moved before this date. The surviving Armenian inscriptions from this decade were all located in the Byzantine sector. They reveal a concerted attempt by Heraclius to integrate Armenian princes and lords into the ramified political structures of the empire through the distribution of honors and other, more tangible benefits. The inscription at Alaman (A.4) confirms that minor figures in Armenia, below the status of tr, were included in the honors system, being rewarded with lesser titles. The inscriptions also show through their synchronisms that Byzantine correspondence was circulating in Armenia during the 630s. The regnal formulae and the eulogizing epithets applied to the emperor were extracted from the protocols of official correspondence, translated into Armenian, and used to locate the inscriptions in time. However, the limited geographical spread of the extant Armenian inscriptions should also be remembered. The literary sources attest a continuing Persian engagement with Armenia during this period. The History attributed to Sebeos reports that Varaztiroc' fled to Taron when Rostom the prince of Atrpatakan sent his brother as darikptetto go and winter in Dvin and to seize Varaztiroc'.207Thus Dvin was still under Persian control in winter 632-633. Moreover, several key Armenian princes
202 203 204

Lazar P'arpec'i, 171.1-178.17; Thomson, Historyof Lazar,?94-?99. Thomson and Howard-Johnston, Sebeos,2:222.

205
206 207

Ibid., 2:223-24. Ibid., 2:227.

Ezr's negotiations with Heraclius over ecclesiastical union: Sebeos, 131.31-132.11. Sebeos, 132.12-35; Thomson and Howard-Johnston, Sebeos,2:229.

72

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

continued to submit to the Persian king; indeed, the sparapetofArmenia, Musel Mamikonean, and Grigor, lord of Siwnik', both fought and died for Yazkert III at the battle of alQadisiyya, on 6 January 638. The inscription at Mren (A.7) needs to be interpreted in the light of Musel's death. It shows that the principal client of Heraclius, Dawit' Sahafuni, held the titles ofpatrikios, and sparapetof Armenia and Asorik'. Collectively these express a broad hegekouropalates, whole of Armenia. The combination of titles is unique, suggesting that the over mony Dawit' had taken over the authority previously enjoyed by Mu'el Mamikonean in the Persian sector and reflected through the title sparapet.The extension of this authority beyond the boundaries of Armenia into Syria is unprecedented. It implies that Heraclius was prepared to make remarkable concessions in his efforts to forge an effective opposition to the Arab invasions after the fall of Syria, one in which Armenian military resources had a leading role to play. In giving Dawitc responsibility within Syria, it is possible that Heraclius envisaged the future deployment of Armenian forces in that region. The concession offered by Mu'awiya to T'eodoros Rktuni in 652, that the cavalry of Armenia would not be summoned to Syria, should perhaps be interpreted in the context of this earlier agreement with Heraclius.208 One other feature of the radical approach taken by Heraclius to the government ofArmenia should be noted. It appears that he promoted men from lesser princely families to It is hard high rank: first Mzi' Gnuni, then Dawit' Sahafuni, and finally T'oodoros kRgtuni. to know whether this was designed to foster loyalty, as each owed his prominence to imperial favor rather than accumulated family prestige, or whether it was simply a reaction to present circumstances. Until 638, the principal Mamikonean prince preferred to operate within the Persian political structure as sparapet of Armenia. Moreover, Varaztiroc' Bagratuni had been compelled to seek refuge within the empire in 633 and had been exiled for complicity in a plot against Heraclius in either 636 or 637.209 The emperor may therefore have had little choice but to use the talents of "new men." In terms of their loyalty, these men have a surprisingly good record once in office, suggesting that Heraclius and those advising Constans II chose well.210 Indeed, the continuity of Byzantine influence within Armenia during the 640s, despite two substantial Arab raids, is impressive and cannot be explained solely in terms of the stationing of imperial forces there.211 Evidently a significant proportion of the 61ite believed that their future lay in relationship with the Byzantine Empire. T'Vodoros Rktuni remained a faithful client until 652, when a sustained Arab assault upon the empire persuaded him that it would be in his own best interests to transfer allegiance to Mu'awiya.2l2But even then, as we have seen, Constans II had little difficulty in reestablishing ties with a large number of princes, reorganizing them under the authority of Muiel Mamikonean.213 It is worth noting that this was achieved through
208
209

Sebeos, 164.22. Ibid., 132.12-133.23.

z10 Admittedly, Dawit' Saharuni was implicated in the plot ofAthalarikos to oust Heraclius, but having seized power, he remained faithful until he was overthrown "by his soldiers." 211 First raid, culminating in the fall of Dvin on 6 October 640: Sebeos, 138.8-139.3; second raid, involving the capture ofArcap'k' and the successful countermeasures under T'"odoros Sebeos, 145.6-147.2. For RiPtuni: the presence of imperial forces within Armenia under the magistermilitumper Armeniamand perhaps the magistermilitumper Orientemas well, see Thomson and Howard-Johnston, Sebeos,2:254-57 and 268-70.
212 213

Sebeos, 164.13-33. Ibid., 165.8-168.39.

TIMOTHY GREENWOOD

73

a high-profile and potentially dangerous visit to Armenia undertaken by the emperor in person, indicative of the considerable strategic importance attached to Armenia and its military resources after the catastrophic losses sustained elsewhere in the Near East. That Constans was prepared to undertake a second extended progress through the Caucasus during the autumn of 659 and that this lasted into the springtime of the following year confirm the priority given to this northern theater of operations.214As the only practicable route for communication with Aluank', Media, and beyond, Armenia possessed a particular significance. Thus it is unsurprising to discover that Hamazasp Mamikonean is presented in the Historyof Atuank'as the principal client of Constans, nor that he was swiftly This removed at the conclusion of the first fitna and replaced by his brother Grigor.215 broke the first link in the chain of clients so carefully pieced together by Constans II, and thereby caused the whole network to collapse. The inscription at Aruci (A.11) confirms that the traditional system of client management, according to which the leading Mamikonean prince operated as the principal client of the dominant foreign power, was retained by Mu'awiya. The preeminence enjoyed by Grigor Mamikonean was expressed through the conventional title isxan Hayoc. This approach to the government ofArmenia would have been familiar to Heraclius, although his promotion of men from lesser princely houses has already been noted. Unlike the earlier Sasanian system of provincial administration, there was no separate Arab governor of Armenia at this time equivalent to the marzpan,nor is there any evidence for the imposition of a new administrative structure. Instead it seems that Mu'awiya was content to entrust the government of Armenia to Grigor Mamikonean, imposing a nominal annual tribute of only five hundred dahekans,if Lewond's report is to be believed.216In the light of Constans II's ambitions for a network of clients, such a limited engagement with Armenia on the part of Mu'awiya might appear surprising, but it is worth remembering the radically altered circumstances in which he was now operating. After 661, Armenia was no longer partitioned
214 Patmut'iwn Atuanic, 183.16-186.10. The passage reports that the visit took place in the "nineteenth year of the reign of Constans" (September 659/660) and notes that Constans wintered at Valarvapat. This is corroborated by an unrelated passage at the end of the compilation which reports that Constans II attended the dedication of the church of St. Grigor at K'alakudaft, founded by Nerses: Patmut'iwn Atuanic, 317.5-11. This refers to the church of Zvart'noc'. 215 Patmut'iwn Atuanic, 181.4-6: "Then the very prudent juanger reached a pact with the Armenian commander and he encouraged him to turn to submission to the emperor of the Romans." Thus juanier entered into relationship with Constans at the encouragement of Hamazasp Mamikonean. Their first meeting occurred in Media: Patmut'iwn Atuanic, 183.16-184.15. Although the passage disparages Hamazasp for his jealousy at the gifts presented to juanSer, its true significance lies in the recording of Hamazasp's presence in the imperial entourage accompanying Constans. Moreover, his jealousy becomes more understandable if one supposes that Hamazasp was recognized by contemporaries as Constans' principal client in the East. The position of Hamazasp became untenable as soon as Mu'awiya had triumphed in the firstfitna. Hamazasp was deposed in either 660 or 661 and replaced by his brother Grigor. For this relationship, see Garitte, Narratio, 405 (the Greek list of kat'olikoi, found in Par. gr. 900), 411 (the Georgian version of this list), and 438-39. A date of 660 or 661 tallies with the time in the office of kouropalates usually accorded to him, either three years (Yovhannes Drasxanakertc'i, 89.19-20) or four years (the Greek list of kat'olikoi mentioned above). Yovhannes Drasxanakertc'i also reports that the kat'olikos Nerses urged Mu'awiya to appoint Grigor as his successor; even if one may have doubts about the likelihood of this request, the involvement of Nerses provides a useful terminusante quembecause he died in about 661. Thus contra Howard-Johnston (Thomson and Howard-Johnston, Sebeos, 2:287), the network of Caucasian clients unraveled very quickly after Mu'dwiya secured his authority. This would also supply the context for Constans' sudden switch to the western extremities of his empire after 661. His plans in the East had been wrecked and there was no imminent prospect of recovery. 216 Lewond, 14.13-15; Arzoumanian, Historyof Lewond, 54.

74

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

between competing great powers. The frontier had moved westward into Asia Minor, relegating Armenia, from a strategic perspective, to something of a backwater. Moreover, the Muslim conquests had been achieved without utilizing the military resources of Armenia. Far from needing to recruit troops to defend against external threats, the principal challenge faced by Mu'awiya was to find ways of channeling the focus of his existing forces away from internal dispute. Thus Mu'iwiya needed only to neutralize the potential threat posed by Armenia as a region of Christian military manpower, and this was achieved by conceding a broad discretion to Grigor Mamikonean. Unfortunately there is no eulogy of Grigor Mamikonean similar to that penned for his contemporary Juangir and incorporated within the History of Atuank217An indication of his power is supplied by the longevity of his regime-twenty-four years as isxan Hayoc'--and by the size of the cathedral church at Aruc. It reflects not only the material resources available to Grigor but also the large numbers expected to attend upon the "great prince" Grigor and participate in reliThe large cathedral church at T'alin should be interpreted in gious ceremonies at Aruc.218 the same way, albeit as an expression of Kamsarakan rather than Mamikonean supremacy. The inscription on the small church at T'alin (A.12) dedicated to the Mother of God confirms that churches continued to be founded at the end of the seventh century. Although the inscription itself does not specify a date-and can be distinguished from all the earlier foundation inscriptions on this basis-this can be determined approximately from a contemporary colophon. The inscription identifies Nerseh, apo hypaton,patrikios,and ter Sirakayew Asaruneac'as being responsible for the foundation of the church. The colophon attached to the Armenian translation of Socrates Scholasticus's EcclesiasticalHistoryhas already been cited in full. The sponsor of the translation is named as "Lord Nerseh apiwhipat patrik, you who are a builder of churches." The titles found in the inscription are mirrored in the colophon. There can be little doubt that the founder of the church at T'alin and the patron of P'ilon Sirakac'i were one and the same person. As we have seen, the colophon is dated securely to the year 695. The plural "churches" makes it more likely that construction of this church at T'alin had been started, if not completed, by this date. The imperial titles accorded to Nerseh are important because they attest an ongoing Byzantine interest in and influence over this northwestern region of Armenia, one that barely registers in the literary sources. The practical advantages for the empire are apparent from a short, isolated passage found in the History of Yovhannes Drasxanakertc'i. This records the victory of "the Kamsarakan and the azatkCofVanand" at the battle of Vardanakert over the commander Ogbay and his forces.219 Although Lewond's History transfers this victory to Smbat Bagratuni, this seems to be the product of later rewriting, and the account crediting the Kamsarakan should probably be preferred.220 This engagement supports the contention that Nerseh Kamsarakan was an imperial client at the end of the seventh century; it also supplies the context for the martyrdom of Sulan Kamsarakan in discussed above. Xar'an/H.arran, When the seventh-century Armenian inscriptions are examined collectively, they do not give the impression of wide-ranging social upheaval. Although the identity of the forPatmut'iwn Atuanic, 172.21-201.15 and 221.3-237.8; Dowsett, CaucasianAlbanians, 109-30 and 142-53. ew erkiwlac Grigor is called barepa't yAstucoy, "pious and God-fearing" (Yovhannes Drasxanakertc'i, 90.2-3) and ifxan mec, "great prince" (ibid., 91.23-24 and 93.16). 219 Yovhannes Drasxanakertc'i, 95.5-19; Maksoudian, YovhannisDrasxanakertcii, 21.1-4. 220 Lewond, 24.2-25.8; Arzoumanian, Historyof Lewond, 59-60.
217 218

TIMOTHY GREENWOOD

75

eign power or the dominant Armenian prince might change, there were no dramatic developments in Armenian social structure. There is, for example, no evidence of any significant Arab settlement within Armenia in the immediate post-conquest period. Indeed, if the inscriptions and the structures to which they belong indicate change of any kind, it is the sudden availability of greater material resources to Armenian princes, enabling them to sponsor such constructions. It is impossible to say whether this was due to the retention of taxes previously remitted to the Sasanian state, to the takeover of land previously under state control, or to the sudden influx of Byzantine silver coins-or to a combination of all three. Far from being an era of unremitting warfare and oppression, the seventh century emerges as a time of stability, prosperity, and material development in Armenia, at least from the perspective of the 6lite.221 Important historical insight into eighth-century Armenia is afforded by the epitaph of Artavazd Kamsarakan (A. 13). It records that imperial honors were distributed to two generations of the Kamsarakan house, implying that Byzantine influence over northwestern Armenia persisted during the first half of the eighth century. Furthermore, the inscription refers to Artavazd as isxan Hayoc'.As discussed above, this appointment does not register in any other source. One explanation for his omission from the standard sequence of appointments is that Artavazd acted as prince of Armenia on behalf of the emperor rather than the caliph. The fact that Artavazd was buried beneath a tombstone in an underground chamber supports this contention, albeit implicitly. Although such a practice appears to have been rare in Armenia at this time, at least in comparison with the characteristic funerary stelae, it was common in the Byzantine Empire. The date of the inscription remains contentious. The epitaph reveals that Artavazd was survived by his wife, Su'an Mamikonean. Although her name does not appear in the literary sources, S. Kogean suggested that she was the anonymous princess who appears in the Life of St. Vahan, prince of Golttn, ordering the saint to leave the monastery in Sirak in which he had settled.222 The colophon attached to this text asserts that it was composed by Artavazd, abbot of Erasxawor, seven years after Vahan's death. Vahan was martyred in "one hundred and eighty-six of the era" and "in the days of Sam/Hisham amir[m]umin,on This combination that side of the river Ep'rat/Euphrates, in the city of Rucap'/Resafa."223 of accurate, incidental detail points to a near-contemporary date of composition. In the light of the directing role played by this noblewoman, Kogean proposed that Artavazd Kamsarakan had died before this event.224Although this argument is ingenious, it relies upon the anonymous "lady of Sirak"and Sulan Mamikonean being one and the same person. On closer inspection, the exact title of the anonymous princess, the "lady of Sirak,"is
N. G. Garsoian, "The Arab Invasions and the Rise of the Bagratuni (640-884)," in TheArmenianPeople from Ancient to Modern Times,ed. R. G. Hovannisian, 2 vols. (New York, 1997), 1:118-25. 222 S. Kogean, KamsarakannerW (Vienna, 1926), 165. For the text itself, see Artavazd, Vkayabanut'iwn srboyn VahanayGolt'nac'woy, Sop'erk' Haykakank' 13 (Venice, 1854). Excerpts from this appear in L. Aliian, Hayapatum (Venice, 1901), 250-53. no. 38.For the date, see above, note 123. Lewond confirms that Sam was the 223 Mat'evosyan, Hisatakaranner, Armenian equivalent of Hishtm: "Again in his place Sam who is Helm rules, 19 years." Lewond, 100.20-21; Arzoumanian, HistoryofLewond, 106. The Life itself records that "Rucap'was the settlement of Sam, the prince of the Arabs" (Artavazd, 35.1-2). This is corroborated by other sources and supports the argument that the date of composition indicated in the colophon is correct. Intriguingly, when traveling to Rucap'/Resafa in 737, Vahan visited Urha/Edessa, where he worshiped p 'rk'akanpatkern, "the Savior's image" (Artavazd, 34.19-22). This appears to be a very early, and hitherto unnoticed, reference to the famous Mandylion of Edessa. 224 Kogean's argument is repeated by Thierry, "Monuments chretiens du vilayet de Kars (III)," 355.
221

76

A CORPUS OF EARLYMEDIEVAL ARMENIAN INSCRIPTIONS

found to be a matter of conjecture. In one version of the text, she is defined simply as iuxanakin,"wife of the prince," and in the other, she is called tikingawarin, "the lady of the district."225 It is probable that the district in question is Sirak, but this detail is derived from another passage. The noblewoman is not identified as a widow or regent; again this is deduced from her actions. Critically, a princess of Sirak need not have come from the Mamikonean house-if anything, this was exceptional. The fact that Su'an elected to define herself in this way rather than in terms of an ongoing political role within Sirak may be significant. Thus there are several objections to Kogean's proposition. It cannot be dismissed out of hand, but it is at best unsubstantiated. It follows from this argument that the epitaph of Artavazd need not have been carved before 737. The genealogy of the Kamsarakan house supports a later date. Nerseh Kamsarakan, appointed isxan Hayoc'by Justinian II in 689 and in power for only three or four years, is traditionally identified as the father of Nerseh, the sponsor of the church at T'alin, the grandfather of Hrahat and the great-grandfather of Artavazd Kamsarakan.226 Although not impossible, it is hard to fit four generations into the forty-four-year period between 691 and 735. It is possible that Artavazd Kamsarakan may have died as late as 760. As has been noted previously, the churches at Zvartcnocc and AruE both carry Arabic inscriptions dating from the second half of the eighth century (A.19 and A.20). As such, they constitute important evidence for the presence of Arabs at or near these sites after 770. The middle inscription at Zvart'noc' also reveals that Ilyas b. 'Isa interceded for his family and his village, implying a property interest of some kind. As has been observed previously, the very existence of these inscriptions raises the intriguing question of who had possession or control of these churches at this time. The fact that two of the most prominent churches in Armenia acquired Arabic inscriptions implies that these sites may not have been under exclusive Armenian control. These inscriptions are also significant because they coincide chronologically with a change in approach to the government of Armenia after the 'Abbasid revolution, one that can be traced through the History of Lewond and the numismatic record.227Both sources attest a growing sophistication in the administration ofArmenia and a greater down-reach into individual districts. The vast northern governorship previously held by Marwan b. Muhammad was divided into several smaller provincial units. Governors were appointed and replaced with increasing regularity.228 Arab officials began to intervene at a local level in the collection of taxation, thereby challenging the authority of indigenous Armenian princes to regulate the affairs of their own territories. The History of Lewond reveals the violent reaction of the princes to these administrative changes. In the course of their rebellions in the 770s, both Artavazd and Muiel Mamikonean are reported to have killed lo225 226

Artavazd,29.16 and 74.10.

Garitte, Narratio, 405, 411, and 440; Step'anos Taronec'i, 100.23-101.16. According to Step'anos, in the year 140 of the Armenian era (691/692) Nerseh was succeeded as prince of Armenia by Smbat Bagratuni Biwratean, who held power for twenty years: Step'anos Taronec'i, 101.16-19. 227 For the numismatic evidence, see M. L. Bates, "The Dirham Mint of the Northern Provinces of the der arabischen Umayyad Caliphate," Armenian NumismaticJournal 15 (1989): 89-111; R. Vasmer, Chronologie von ArmenienunterdenAbbasiden, von al-Saffachbiszur KrdnungAschots Statthalter I., 750-887 (Vienna, 193 1), 1032; X. A. Muietyan, "Abbasyan dramneri t'olarkum= xalifat'i 'Arminia' nahangum VIII-X darerum," IFZh (1974) 4: 143-56 and appendix; M. Bonner, "The Mint of Harfinabad and al-Haruiniyya, 168-171 H.,"American NumismaticSocietyser. 2, 1 (1989): 171-93 and pls. 14 and 15. 228 For the standard list, see Laurent, L'ArminieentreByzanceet l'Islam,app. 4.2, 408-33.

TIMOTHY GREENWOOD

77

These inscriptions fit cal tax collectors, operating in Sirak and Bagrewand respectively.229 neatly into the context of greater Arab involvement in these districts. Having exploited these Arabic inscriptions in support of the view that the cAbbasidapproach to the government of Armenia was far more intrusive, it should not be forgotten that those from Zvart'noc' also display several highly unusual features setting them apart from all other contemporary Arabic inscriptions. Although the exact wording of two of the inscriptions remains unresolved-a consequence not of their poor condition but rather the atypical nature of their content-Hoyland has securely identified two remarkable passages. The first inscription refers specifically to the governorship of one Yazid b. Jarad and seems to correlate his tenure with the year A.H. 154. Although it is tempting to identify this individual as Yazid b. Usayd as-Sulami, governor of Armenia on at least three occasions during the eighth century, the reading does not support this.230Such a synchronism is unprecedented in Arabic inscriptions. We have already seen, however, that seventh-century inscriptions in Armenia employed similar synchronisms. This Arabic inscription appears to be reflecting Armenian practice. In support of this contention, it is significant that the relevant passage identifies a second individual after Yazid. Although his name cannot be read with confidence-it may be Bgr, Vgr, or another variant-it seems that this is because it is an Arabic rendering of an Armenian name. Thus the inscription is dated by reference to the hijra, the governorship of Yazid, and the headship of an Armenian-a unique combination of elements in an Arabic inscription. Furthermore, we have already seen how Ilyas interceded for his whole family, his house, and his village. The startling conclusion is that these inscriptions represent a fusion of Arabic and Armenian elements. On the one hand, their language is Arabic, they employ hijra dates, and at least two of them invoke Allah; on the other, the synchronism and the scope of the intercession both sit very comfortably within the Armenian tradition. The marriage between al-Djahhaf and one of the daughters of Mu'el Mamikonean after the latter's failed rebellion represents another expression of this process.231 The latest Armenian inscription included in this corpus (A. 14) also attests the changed conditions experienced within Armenia, albeit from the other side. The inscription records a small-scale, practical improvement undertaken by Uxtaytur and his brother rather than a grandiose extension, rebuilding, or decoration. Its limited nature seems significant. Moreover, the inclusion of the curse against anyone who destroyed the newly provided water course indicates a fear, and perhaps even an expectation, on their part that this would occur. Again this implies ongoing disturbance in the district. The tone of the inscription is supported by an otherwise isolated report in the History of Yovhannes Drasxanakertc'i recording a massacre at T'alin involving the killing of seven hundred people and the capture of one thousand two hundred.232This occurred during the eight-year tenure of Sion as kat'olikos of Armenia, ca. 767-775. Finally, it is worth recalling that this is one of only
Lewond, 138.12-14; Arzoumanian, History of Lewond, 129: "[Artavazd Mamikonean] arrived in the province of Sirak, at the village of Kumayr, and he killed the assessor of tax (hramanatarharkin)."Lewond. 139.18-21; Arzoumanian, Historyof Lewond, 130: "[Muiel Mamikonean] went into the province of Bagrewand and with him 60 men and having seized this collector of the tax (zpahanfolsnharkin),whose name was Apum'ur, and those who were with him, he struck and put them to death with his sword. And he interrupted the collection of tax from this land." 230 Laurent, L'ArminieentreByzanceet l'Islam,app. 4.2, nos. 27, 28, 29, and 33. 231 Ibid., app. 3, "Les principautes arabes d'Armenie," 383. 232 Yovhannes Drasxanakertc'i, 107.7-10; Maksoudian, YovhannisDrasxanakertc'i, 23.17.
229

78

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

two Armenian inscriptions to date from the eighth century, the other being the epitaph of Artavazd Kamsarakan (A. 13). This sets into even sharper relief the prosperity and relative security enjoyed by much ofArmenia for most of the seventh century and attested through the multiple church foundations.
CONCLUSION

The small size of this corpus belies its significance for the study of early medieval Armenian history. As authentic contemporary statements, these inscriptions afford important historical insight. Their value extends far beyond the naming of individual sponsors or providing exact dates for the construction of certain churches, important though these are. They attest something of the motivation behind such foundations. They confirm the sequence of chronological mechanisms employed in Armenia in the course of the seventh and eighth centuries. They operate as independent controls against which the social terminology found in the written sources may be tested. They reveal that members of the Armenian 6lite could be defined in more than one context at any one time; thus the founder of the church of the Mother of God at T'alin is identified as Nerseh apo hypaton,patrikios, and terSirakayewAsaruneac'.Through the use of synchronisms, the inscriptions also reflect the political hierarchy that existed within Armenia during the 630s as perceived by a contemporary. This comprised three tiers: a single, dominant prince serving as the principal client of an external power; individual lords exercising power at the level of the district or gawar; and those who enjoyed a lesser, subordinate responsibility within the gawar, probably at the level of the village. The inscriptions also supply vital information about the conduct and capacity of noble-born women. Finally, they attest a bold initiative on the part of Heraclius to attract Armenian lords to his service through the distribution of imperial titles and attendant material rewards. A parallel initiative in the ecclesiastical sphere is attested by the inscriptions in Greek commissioned by the katcolikoi Ezr and NersEs III. Successive emperors recognized the strategic potential of Armenia and continued to confer titles, at least upon those lords who controlled districts to the north and west of the Araxes valley, up to the middle of the eighth century. This pattern of contact is largely ignored within the surviving written sources. The Arabic inscriptions at Zvart'nocc(A. 19) and Aruc (A.20) attest the change in approach to the administration of Armenia which took place after 760. Their very existence suggests direct Arab intervention, if not settlement, at the level of the district. The unprecedented style and content of two of the Arabic inscriptions may represent a synthesis of Arab and Armenian elements. Oriental Institute, Oxford University
I should like to express my thanks to the two anonymous readers for their valuable observations and suggestions, to Professor R. W. Thomson for his comments and his translation of the Georgian band inscription at Samgvilde, to Dr. R. Hoyland for his translation of the Arabic inscriptions, and to Charlotte Rouech6 for her willingness to share her epigraphic expertise.

Appendix 1 A Corpus of Early Medieval Armenian Inscriptions

The inscriptions have been transcribed in accordance with the Leiden Convention. [UP] {UP <UP> (UP) ..(P [. . .] [-] Letters restored by the editor as once having been inscribed but which are no longer preserved Superfluous letters added in error by the engraver and excised by the editor Letters added by the editor which the engraver has either omitted or for which he has by error inscribed other letters Letters which complete words left in abbreviation in the text Letters of which sufficient traces remain to print them in the text but not enough to exclude other possible meanings Lost or illegible letters equal to the number of dots for which no resolution is proposed Lost or illegible letters of an uncertain number Start of a new line on the stone where the text is not printed in the same configuration as on the stone

As far as possible, the transcriptions follow the layout of the original inscriptions. Each line of text has also been numbered. The band inscriptions have been set out differently, with each exterior face being allocated a letter in alphabetical sequence. The inscription at Naxciavan (A.8) is the one exception to this, the letters representing the sequence of individual blocks. A. ARMENIA

Inscriptionsin Armenian
1. Tekor: Late fifth century [478-490] and unknown (Fig. 1). In architrave above tympanum of western entrance. (Venice, 1881), 132, whence reCopied and published by L. Ati'an, Siraktetagrutciwn patkerac'byc K. published by Kostaneanc', VimakanTaregirC'uc'akotovacoyarjanagrut'eanc' Hayoc'(St. Petersburg, 1913), 1; photographed and published, with full bibliography, by G. Yovsep'ean, Gr 'ut'eanarueste hin Hayoc'mif K'artizhay hnagrut'ean(Hay gri giwti 1500 ameakiart'iw) (Valar'apat, 1913), no. 2 and pl. 2; published by N. Marr, "K datirovke ktitorskof nadpisi Tekorskogo khrama," KhristianskifVostok 3.1 (1914): 56-71, whence republished by A. A. Manu('aryan, K'nnut IV-XI dareri yun Hayastani ?inararakanvkayagreri (Erevan, 1977), 39-40; separate photograph published by S. Der Nersessian, L'artarminien (Paris, 1977), fig. 32; separate photograph and squeeze published by M. E. Stone, D. (Aarhus, 2002), 14, ill. 2 and 15, ill. 3. Kouymjian, and H. Lehmann, AlbumofArmenianPaleography

80

A CORPUS OF EARLY MEDIEVAL ARMENIAN

INSCRIPTIONS

1(5). [ ]U-U{}<fl>flUP

bb

SlUBPlbb SOlRPfRl LU'uLI PIh8blb

? RP;[-]

U b LUU8,111

I BU UU MPU'GUL 5(1). UU4MUQ 9U1lUIZPb P h118 Sahak Kamsarakan built this martyrium of Saint Sargis for his intercession and of his whole family and wife and children and loved ones and [-] And this site was founded by means of Yohan of the office of kat'olikos of Armenia and Yohan bishop of Ariarunik' and Tayron elder of the community of Tekor and Manan hazarapetof Uran Hor*om[-] 2.1 Hiip'sime: After 616/617 and before 628 (Fig. 2.1). Located on western elevation.

,RARl' {fl u 2(4). 6UP b UP BPRIU uL{L}<<l>lhPlAUlh' RtlIL 4bWl~ULPUIhPSkblU h 904U110 31)U '141kU.PiLL 1U1L( 3. BPPP PUP61LPbU11PAPb', L UPPfi1P b u?9bPU UP l-IhUtlihPbb P rflPk-UU UP bPhU lUOl.U 4(2). UPPrO1 lULPbUPh

4l R lP90qPSh AAhPlbflU

kat'ulikeEfmiacin ew hing gawarac'nAraCopied and published by Y. Sahxat'uneanc', Storagrut'iwn vols. (Ejmiacin, 1842), 1:264, whence republished by L. Ali'an, Ayraratbnasxarh ratay,2 Hayastaneayc (Venice, 1890), 238; copied and published by G. Ovsepian, "Komitas kat'ulikosi mi nor arjanagrut'iwn," Ararat 10 (1898): 442, whence republished by Kostaneanc', VimakanTaregir,2; photographed and published by Yovsep'ean, K'art-z,no. 3 and pl. 2; copied, drawn, and published by I. A. Orbeli, Izbrannyetrudy (Erevan, 1963), 407; published by A. B. Eremian, KhramRipsime (Erevan, 1955), 26, whence republished by A. G. Abrahamyan, Hayoc'girev gr('utcyun (Erevan, 1973), 83, and and 79; Stone, Manuc'aryan, K'nnut'yunHayastani, photographed published by Kouymjian, and and 1. 112 Lehmann, Album, Inscription

1. 2. 3. 4.

I Komitas sacristan of saint Hr*i{w}p'sime was summoned to the throne of saint Gregor. I built the temple of these holy martyrs of Christ 2.2 Hiip'sime. Located on eastern elevation, interior. Copied and published by Ovsepian, "Komitas,"442, whence republished by Kostaneanc', Vimakan 2, whence republished by Manuc'aryan, Kcnnut'yun Taregir, Hayastani, 79. \ IU 1. "r(PUSIUr) ( aUS -u) tr BfOc9 ?F.9A1AlbS 2. ZILYII rrUf P'i PArkQ(.U 3. U~b UPPRFIT Zl4PU'WU Christ God, remember Komitas kat'olikos of Armenia, the builder of saint Hiip'sime 3. Bagaran: Completed 8 October 629 (Fig. 3). Single-line inscription at roof level, beginning on northern section of western elevation, running across western, southern, eastern, and northern apses and elevations, concluding at western face of northern apse. Copied and published by N. Marr, "Imia But ili Bud v armianskof nadpisi VII-go veka po R.Kh.," Obshchestva 7 (St. Petersburg, 1892), 322; copied ZapiskiImperatorskago RusskagoArkheologicheskago and published by Orbeli, Izbrannye trudy,391, 393, whence republished, with separate photographs,

UP U- 7bS!lU )WtSU U'Ol9.' i ' BUP 2U0(1B PO? EU2PU I.b{b } INik UlPPROI8 hi;'Plbb 9P-P9r./PP RSllUlUP l~U8PltUUP(PhUSAIU)P UPPRI.i .

NOTE:The sequence and numbering of the figures reflects the sequence and numbering of the Corpus. There are no photographs of the Armenian inscriptions at Htip'sime (A.2.2), Bagavan (A.5), and Naxiavan (A.8); the Greek inscription at Mastara (A. 15); or the Arabic inscription at Aruc (A.20). The photographs used to illustrate the Armenian inscriptions at Bagaran (A.3) and Alaman (A.4) were not taken for this purpose and reveal only a fragment. They remain, however, the only published impressions of these inscriptions, which are long since destroyed.

1 (Inscription A. 1) Tekor (after S. Der Nersessian, L'artarmenien [Paris, 1977], fig. 32)

2.1 (A.2. 1) Hiip'sime (after G. Yovsep'ean, Grfutcean aruest khin Hayoc'mgj.KiartJ& hay hnagrutcean (Haygri giwti 1500 ameakiaitciw) [Valargapat, 1913], pl. 2 and no. 3)

3 (A.3) Bagaran (afterJ. Strzygowski, Die BaukunstderArmenier undEuropa,2 vols. [Vienna, 1918], 1: fig. 34). Although of poor quality, this photograph confirms the location of the inscription across the top row and Orbeli's readings of facesJ and K.

46 O
.1 p

b-)

,? , .: -, " :r-.
:) Sb

?,.<

.. .

?..

?xf,
'L41C

Tow,

00
iA

qet~u~??,1 1 """"""""""""""4
Mot

'

' " "ic:

off? 4.0 .". .do

4 (A.4)Alaman (afterStrzygowski, Die Baukunst, 2: fig. 602). The photographconfirmsOrbeli'sreadingsof faces H and I.

4 1 ?

;1'

?~ ~;~

..

t i

b?

I~
n.r

C
~ ?. ??h ,
r

,,

?,

f?l~
r, ?E

t, C~i

u 13 r

?? I

R' ~ u ~ rf .;V ~ ;u~ ?y??'jC

ti

t'Sc~

?~jc~

~?

-?i~

?4 ~r

c~ ~? _, ?rS,

ew vimakan 6 (A.6) Avan (after K. Lafadaryan, Erevan.Mfjnadaryan hularjanner? [Erevan, 1975], pl. 139) arjanagrutfyunner

1 IjIN, TP ~

-r.:

a.,

~-%4t -

~ ?'Pi~f~P1~,ON LA

7 (A.7) Mren (after Yovsep'ean, KIartiz,pl. 3 and no. 6)

;WOO 7L?

:~~~?`46

, , 'A -s

?' ir

??57 : Cx

?~dIL

rd-

gAT .rit

9 (A.9) Zvart'noc' (after Strzygowski, Die Baukunst,1: fig. 31)

,?

IJ;` r? ::~:
too

.0

IfI

0,,
.I`.

'.

,?~

? .ic,

LY~~

10.1 (A. 10.1) Mastara (after Yovsep'ean, Ifartez, pl. 4 and no. 7)

i' ILLI t it'- 'l'!

._OIL

r",-'

." ./

Rl

.t " . ii -":
, . . -

10.2 (A. 10.2) Mastara (after Yovsip ean, KIart~z, pl. 5 and no. 8)

~-t? ~t
Ntv

?s

vt

2I.
b4k

cA s-v

,; tAI g\ Ir

7 vols. [Zug, 1980-90], 7: microfilm 159, C8, and 10.3 (A. 10.3) Mastara (after V L. Parsegian, Armenian Architecture, A. with the kind permission of Dr. V Parsegian)

r? r . ~. ''?H1

.i ,tr~t~t~:' (( ~~(
I ?,
\?

C ;,Y. ~

~~~tlj! ~L?'~C rYA~ ~4?F~

~i~IJ(

.. H L J r_, ,, .x ?;r r ~iil * '

7.

Wt
iI

~Sdlr ~S~r :

"r j~~lrl~c.n, c-rt '

;3P
~'%

.G -r:i1.
~'6', ?r rI -1rtt\' 1~ 1

7: microfiche 159, C9) 10.4 (A. 10.4) Mastara (after Parsegian, Armenian Architecture,

40? .Ylow

41-4b
i'

%.
.-.

?~IL

?;Ca

11 (A. 11) Arue (after Yovsep'ean, KartJz, pl. 6 and no. 9)

i,, c 'g '

t L.i r.' . *

?. ,
~ % t: r 4 r ,-c

-i -? ?

12 (A. 12) T'alin (after I. A. Orbeli, Izbrannye trudy[Erevan, 1963], pl. LVIII)

S""4

U I) Ul3 $ ,TPl, I,',,. ,II?,'. IY fr FJ

?

h
f

l n q . ?
-u

U A ' JU PUO P fUt D It-.

. .

r UU ! S ' ru.-U 7r LuJ ,j ,,.'

""

L ..

.f t.I ",.
.._ ..

0 tJ !3tu
. , _ _" . .__:._~-;~

. .. b ... l

13 (A. 13) Naxiavan/Noramanuk

(after Yovsepcean, Kartiz, pl. 5 and no. 10)

Ap,;

Cl!Im

O~~~s

'-,
N ..

, -..

-..

--

.,41

14 (A.14) Tcalin/Uxtaytur(afterYovsepcean, pl. 7 and no. 11) KIartiz,

u r I $--~ r'

v. ..rC? ~h~.r, *i? J S .? c o?

'Z I;?I r I ?\ .L iG i. 'Y u .;?r ~ ??* L?

I,

ri~j H:i`f. u C ~~?' r r? irr ?~b

J~a`~P;~E~dC~P rPsCR~b~g~ ~yj

~i~F .z~(W "P Ti

~-.-~~"a~B?

~L;~-?ll

-i~

-YYA

)fB

:?CT; i L. LU* (.? r;

r t.: r ~i~?gW-

* ?-Y

~t,
J
LA

,~:~'~rS~S~B~PS~l~kl*~~;?~~ r
C re r . ~?f ,?rC?? i ?: ?$ j 9~h~?z? r

.,
k

J r

~?,

r j r

:~h
L 4' c J Il C .,

?I "`

i; Il~:i

~Slfi~t'
c ~

i,:

?-

~...` ?

t,

P 16~ Ipr. rt~~l o ?? i?4j .r i C

O
i ?- PI )I .a~c ?~p

~bc~! r*($
,.I ?. J:,, ;rc~

?I) r

~~ w~w~

a:
..

.c

~-~Fi?:

?-

j: c. c

???? c? c. * A -c ? I

?~,?~C P L: ~p?

1
..

i' t?'?
? d! r ~b~Le ~II~. u: 'C

?,

I.. f, ~C'~ C/

u~,~u~ ?~

:~d~.~f~ ? .. ~)~??r?i ~cv~a?~lp~r ~7n ~ .%rJr:-~,a~ce I? ~c

i"?I

c ~~ X

~o
~L~~h: (k ~C~rT~

r .4.hc;~?rr~9~

o 7tfi ~ )1)

I r'','' ' "~

P * c )U .c.f

i~

;rl, c?E ?5 ?~ s

~~??

r~

*E a~.

r?? .? (kA: # :~ki i

16 (A. 16) Ereroyk' (after E V Shelov-Kovediaev, "Zametki po grecheskoi epigrafike Armenii," IF h [1986] 1:60)

Ir?
moo? 0

I. All A A

17 (A.17)Ejmiacin(afterParsegian,Armenian 1: microfiche8, B 10) Architecture,

**4

?,o

10-.

te
.Jr

18 (A. 18) Zvart'noc' (after Strzygowski, Die Baukunst,1: fig. 30)

bft
10

*f:

0-O -' IL

A.. of-

_tw

19 (A. 19) Zvart'noc' (after A. A. Xac'atryan, Korpus Arabskikh Armeni VIII-XVI vv. [Erevan, 1987], pl. Iv and nadpiesye no. 6)

"raft
rfV.1 w. ..? 1 tvf., w Jl jl.%%J, AP*?

..ov.
--4'0* illro~ W410 ,6% jl

lot.
A 4,11111.

.;:
4o

w-r
" :c

20 (B. 1) Musrara (after Der Nersessian, L'artarmenien, fig. 45)

'10
40 OP

~L~Y-C~. ~?~'~Clli~-?~~L.U -`~;~;?~-~~t~'P ??;~ '?'?~x ~?~t~~ ~=r~~7ir~c "?low. ~

??
Nor

??ai.

00

IT 001 N

4w~

bC

lie's iiowl

21 (B.2)Mount of Olives/Su annan (afterE. Loukianoff,"Le Musee du couvent russe du mont des Oliviersta Jerusalem," BIE 13 [1931]: pl. vII)

i;;;:

v; dr

01

C
Stwe

"?C'

"IJ
nrr
. ,

22 (B.3)Mount of Olives/Jacob (afterLoukianoff,"LeMuse'e," pl. v. 1)

.d,,3 $?EIt ?lb !14 a

0t~'

(?

~JP

.".4Z

.~

V-Ar

ri: 3-f;u? 2Eko lp ,*

23 (B.4)Mount of Olives/Valan (afterYovsepcean, pl. 1 and no. 1) Krartiz,

C.

.0:
;; ,

--A,, ,,-NN S,

L:
,. . .',,
.

;I:IQ
x .-,

)6..

4,

\r~r

~rK

'0VP ?? ; 1c~? J~?

4.44

-re

as

, "io WRW

.4.

Review22.3 [May-June 1996]: 24 (B.6) Musrara/Ewst'at (after G. Avni, 'Jerusalem as Textbook," BiblicalArchaeological

40)

TIMOTHY

GREENWOOD

81

by J. Strzygowski, Die BaukunstderArmenierund Europa, 2 vols. (Vienna, 1918), 1:33, fig. 34 and 85; also republished by Abrahamyan, Hayoc'girev grc'uttyun, 79, and by Manut'aryan, K'nnut'yun Hayastani, 59, 62.

A. B. C. D.

E. U-0rU 18U] URAhPP O-4b.b[h

G. H. I. J. K. L.

UPUUlhP'i Ub 21rPP1 /UP P, Il KF f,)1u8j HILh P PaULUhS 2ZP tIP PthS lfhbL291 U.ORU

I 9PP bUP Ph't t 0~A hIulhJ'IL~ fhIn', ~/rflPfbUhlP', qPF1hS l UlI'U l'uLftPIu U6Ph ' h gbUS 6PSh' USULPMl Ul lhPPP bU-Ub-hUlU IL SP PShb LU`LPUhb 7'5108t P ? b U'UU
I 4LPUShPtS

M. [-] N. [-]

O. [P (fIr]P{t(r})<i>l',tP9lA

s. T. [Uc][--] [--] PL't, R UPgUP 4 U. nrPOh t Fry1UU4PUcth UZU rnChU V. 8 ULI'tJWPAp 94/PP 'ULt3USkU6R LBUPIRPfbU 3tUS/P-P ULU
W. 4P %[... .]I'4G ZPZJL{P}S UP SPULPPI/
UP IPl/uSIh X. f
8hOtU IL(UShlPU)L)

L P. P ZGI7 ILU"UbShP GUbhol SUPt/PfdOUl-i (4L'UJ Pb 4 Q. UIL' R. PO /MPFUPLPR

UUl/lrP 1U/LUUP itPAUt, {d-{J}<i>1 IP8/

In the thirty-fourth year of king Xosrov, the blessed Lord But Aruelean set out the foundations of this holy church. In the thirty-eighth year Gobt'i and Xumat' killed But and three years after the death of But, Annay the wife of But completed this holy church in the month Tre, on day 20 [-] in the office of marcpan(= marzpan)of Varaztiroc' aspetof Armenia, in the lordship of Vahan Aiielean, in the (spiritual) oversight of[-] brother's son Ewserk. May God remember and have mercy on the daughters of Grigor Vanandac'i, the husband ofA'anu' Kamsarakan, Annay and the little children G[. .]han, Hrahat and Tiarwand [and] Susan, princess of Sahak Rahasi 4. Alaman: 5 October 636/4 October 637 (Fig. 4). Single-line inscription on fourth row of blocks below roof level, beginning on the west of southern fagade, extending across southern apse, and continuing to northeastern face of eastern apse. Copied and published by Aliian, Sirak, 125, whence republished by Kostaneanc', Vimakan Taregir, 1; copied and published by Orbeli, Izbrannyetrudy, 395, 397, whence republished, with separate photographs, by Strzygowski, Die Baukunst, 1:36-37, figs. 184 and 187; 2: fig. 602; also republished Hayastani, 89. by Abrahamyan, Hayoc'girev grc'ut'yun,85, and by ManuE'aryan, K'nnutjyun 9 t A. [1f]UuL', UP U9'tbUWhRt UPUIlh S lNULULRPt UP "UPUthP PPU UPbULUTLPPPbt B. PULUL98 C. I8 SUILILbUP U D. tO

82

A CORPUS OF EARLY MEDIEVAL ARMENIAN

INSCRIPTIONS

E U6-i W PhBll U1Uqn"murk[b] &U G. P 6ftlPhUSP6UP UrIPhILU hUW

H. Ah' ghltgSU 1 1uR

E. hl. nU[ h I]uPrnP

I. hPP U, 4Y[/],sIJ J. U li U'11 Ur'P 411lPf1

In the twenty-seventh year of Heraclius pious king in the time of Nerseh lord of Sirak and AMarunik' and of T'eop'ilos bishop ofA arunik' I Grigor etustr and Mariam my wife we built this holy church for the sake of our souls 5. Bagavan: Begun 15 June 631 and completed August 639. Single-line inscription at unknown height, extending across southern, eastern, and perhaps northern fagades. 2:591; copied by Ali?an, Ayrarat,529, whence Copied and published by Sahxat'uneanc', Storagrutiiwn, Vimakan and Kostaneanc', 1; Taregir, copied republished by published by Orbeli, Izbrannye trudy,386 and insert, whence republished by Abrahamyan, Hayoc'gir ev grc'ut'yun,84, and by ManuiC'aryan, K'nnut'yunHayastani, 70.

UP U'P [bW]{rJ}h'I, A. -UUIJ-L (dhUULARIfiPP U&PWIQP ZP[i]ShW UL(USPII)WU1109114

Ufl[UR]8t/'

[1P 1]tP OPOU1Rb eP U [-]

UbP [PPrl(]AL c. sUr40UPflU S, '6 P-h E. [r6 UP ]PPfIh UPh

[-] t(USIIrTI)8 4g11, P4-I [.18] .Agdfrhbfh/,

] E htPP 9u. hfir -f U'UPc '6tLPLULPP ,P [,L .Z1

fl]

H. fiU SOQhfltlU ?0 '6b[ULUAP] 141U'TLh ILU'6Ut'6 PUPh P4 I. LO~1SUflPR.P J. lU'P PUP Pt/li P,&

K. t Qf [d]6-0,o

In the twenty-first year of king Heraclius protected by God, in the month Hrotic' which was day thirty, I ... elder of the community through the grace of God made a foundation of this holy church and in the [twenty-nin]th year of the same king Heraclius in the month Nawasard, which was day [-] the day Friday, there was a completion of this same project of this site through the expense, through the will, and in every matter, through the instruction of Israyel GoralEec'i 6. Avan: Between 630 and 640 (Fig. 6). Carved on three blocks which were originally positioned on eastern elevation, at an unknown height. K. Lafadaryan, Erevan. Mijnadaryanhularjanneri ev vimakan arjanagrut'yunneri(Erevan, 1975), no. 194 and pl. 139, whence republished by V. L. Parsegian, ArmenianArchitecture, 7 vols. (Zug, 19807: microfilm A6. 159, 90), 2. PPS 8{<^}<oi>XKo;S?toe0 UUW<fP>Ij fllu 3. UA fEz ks(PaUSU) Lord Ezr kat'olikos of Armenia true servant of God Samued servant of Christ

? represents lBRItfi and CfMUfl'l, Note that The cryptogram is mathematirepresents UrLPUI. 2j*.. cal. Each of the original thirty-six letters has a numerical value. Those letters which represent nu-

TIMOTHY

GREENWOOD

83

merals 1 to 9 have been exchanged for those representing 1,000 to 9,000, and those representing 10 to 90 have been replaced with those representing 100 to 900. 7. Mren: After January 638 and before 11 February 641 (Fig. 7). Three-line inscription, carved onto a single row of blocks, located immediately below roof level, on western facade. and whence Copied published by Alilan, Ayrarat,114, republished by Kostaneanc', VimakanTaregir, 2; photographed and published by Yovsep'ean, K'artez,no. 6 and pl. 3; published by Orbeli, Izbrannye trudy, 397, 401, whence republished, with Yovsep'ean's photograph, by Strzygowski, Die Baukunst, 1:41,43, and fig. 38; also republished, with new, partial photograph by N. and J.-M. Thierry, "La cath6drale de Mren et sa d6coration" CahArch21 (1971): 44-45 and figs. 1 and 2; also republished by ManuE'aryan, K'nnut'yun Hayastani, 64; separate, partial photograph published by P CuArmenadal quartoal diciannovesimo neo, Architettura secolo, 2 vols. (Rome, 1988), 2:815.

1. [--]

2. [8h S4L0B]]

nfP/hP Plzqh fldCl4PlhPP PlP9ulOq.) . f/.4.'lUr~fflAb ] [rU]/UA/-nl9,4 ',f7P1P#1U rsi U"IPU[NU] [P]P 4UtSr1pr

[--1

bU1 O6h unrPP

0 0U4fn 8UP finUlAP

U. -] ,[.] UP[ .] [J[-

]6ihhU

UPb

3. [JPR]'98U6 U A- ChtPhU(I UI)PPP bbIPigu SUL UPHr Ut'JB8[PU P PULPJUV4'4U]AiPkPhhP IL & u1 P WbrUL1P 'iu wru'Pirn WUP[....1 6su.
[-] of the victorious king Heraclius, in the office of prince [-] of the all-praiseworthy patrik, kourapalat, and sparapet[of Armenia] and Syria and in the office of bishop of [-] [T'e]ovp'ilos and in the office of tanutir of Nerseh lord of [Sira]k and A'arunik', this holy church was built [for the intercession] of the Kamsarakank' and Mren and Aser [ ... .] 8. NaxEavan: Between 630 and 650 (?). Incomplete single-line inscription, perhaps on fourth row of blocks below roof level, extending across northern and perhaps western elevations. ew i Mec Hays (Venice, 1864), 215, whence Copied and published by N. Sargisean, Tetagrut'iwnk'P'okr republished by Ali an, Sirak, 136; copied and published by Orbeli, Izbrannye trudy,431, whence republished by ManuE'aryan, K'nnut'yunHayastani, 90; single, poor photograph published by Strzygowski, Die Baukunst, 1: fig. 36

A.
C.

[--]U B. SIPWUi8
D. U64-0 E. U 1BU
E G. H. I.
WU /lr 1US

ThIL?IL.HA 1*&8UIUP 'tI~U

F'uPSh[PI] J. Ull RPa K. .UHLU tL1 L. U~LL[Pl]U M. L BU1PU IL [U] N. POU

o.

94/at, 'tr~

Q. ILP~ 'i

84

A CORPUS OF EARLY MEDIEVAL ARMENIAN INSCRIPTIONS

R. OPUUbZ S l(W
S. P 4h 4hP

U. 8 [t], [--]
[-] b alu'4llUILrPSh[PR]UL'U USPgULLPbIS1U iN USboUZIl'U UHLR/UR[1 Uw'itRoU fIPl89 RU.2llU 4 s ]h [-] P, ')uSP4h _j1UP6UhU pb -Ph [] ruuiPr ,1WIII[PL]lrwb uPPutg[il .Phq9 jh(L.b

[-] worldly ones, O Step'anos apostle and protomartyr, you for whom the parent of Grigor adorned this martyrium, in the time of Nerseh patrik of Sirak and [-]

9. Zvart'noc':After 644 and before 660 (Fig. 9). Located on sundial, unearthed in course of excavation. Photograph published by Strzygowski, Die Baukunst, 1: fig. 31; photographed and published by T'. (Erevan, 1942), 254, whence republished by Abrahamyan, T'oramanyan, Haykakancartarapetutcyun ev Armena, 2:812. Hayoc'gir grc'ut'yun,85; separate photograph published by Cuneo, Architettura

I L[ths](Iv) [W]Ux u.I I JU kU[U] u[mahtlol

PJu.h unflrcP]

f!V)4/'jU

May they pray to the Lord all-holy the one who in time receives 1234567891011 12 10.1 Mastara: Mid-640s (Fig. 10.1). Carved onto single block positioned directly above window over entrance into southern apse; extends onto window frame. Copied and published by Sahxat'uneanc', Storagrut'iwn,2:46, whence republished by Kostaneanc', VimakanTaregir,3; photographed and published by Yovsep'ean, K'art-z, no. 7 and pl. 4, whence translated, but not republished, by Strzygowski, Die Baukunst, 1:44 and fig. 39; also republished, without photograph, by ManuC'aryan,K'nnut'yun Hayastani, 53, and with photograph by Parsegian, 7: microfiche 159, C5; separate photograph published by Parsegian, ArmenArmenianArchitecture, ian Architecture,1: microfiche 35, C4.

2. 3.

1. 8ulrU fnPntPnU[-] S(OU1t), hSU

lbfllhbI

4. I?LhPUl?%JPh-?P1ZPU 5. 1rPd-#IL'

{l}<>9lghb L(USIIU)I)[UBPh]i, SIPL'U P,'iSUHUP U16

P ThPsnr9U U 6. [l](rPhusf)VJ 9T4ILh Ur hUPnP l P6hh2 4I6[... L(UsnIh).? 7.

P[--] In the years of Lord T'edoros bishop of Gnunik' this holy house was built to expiate the unworthy Grigoras. Christ God, be compassionate to Grigoras sinner and to me Kep'[...] and [-] 10.2 Mastara: Mid-640s (Fig. 10.2). Located in upper register of southern section of western elevation, above level of window; however, lines 6 and 8 are carved onto southern face of western apse, perpendicular to the other surface. Copied and published by Sahxat'uneanc', Storagrut'iwn,2:47; copied and published by AMilan, 3; photographed and published Ayrarat, 135, whence republished by Kostaneanc', VimakanTaregir, by Yovsep'ean, K'arthz,no. 8 and pl. 5, whence republished, without photograph, by ManuE'aryan, K'nnut'yun Hayastani, 54, whence republished, with photograph, by Parsegian, ArmenianArchitecture,

TIMOTHY

GREENWOOD

85

7: microfiche 159, C6-7; separate photographs published by Parsegian, ArmenianArchitecture,1: microfiche 36, B3-5; additional photograph published by Cuneo,ArchitetturaArmena,2:814; photographed and published by Stone, Kouymjian, and Lehmann, Album, 113 and Inscription 2.

1 P 1. +9CUlWtll" 'S .R(USflh'rf)9 tPS 2. 941PPRUL8 U Oh l1S60U UhbhlbUh P 9q 3. lIl fp fPh/fRPh ABu Ihfl.g Pl 4. 0hPt't1. S .[rl[UlPP]l1UHU UlRA1Lh 0

2 9P

6. UP.R[P] lt / Ut

5.

U4 hU [..] [69]hg0lflU

10. t[ B R]lO

.h[t UPHIIIJ ' 181e1LUPt 7. UP U lZltAPIrLP1L P 8. WOUfItPIh uShillS 9. WIUPui 2 U Oh flHP tjlhlaP Ph26 Jrh[l[]
GhmBaSt

i !

W ]LL?/H[P]

n/6AU

[,]

I thank God who permitted me Gr4egorasSiwni and beloved nephew Grigor to build a house of glory and through this made me ... bishop ofApahunik'. This is a refuge for Mazdara (= Mastara), a place of prayer for the faithful, a place of expiation for sinners, a memorial for me and for mine. And you who pray, remember us 10.3 Mastara: Mid-640s (Fig. 10.3). Located on two unequal blocks above the window which is itself positioned above entrance of western apse. Copied and published by Alifan, Ayrarat, 135, whence republished by Manuc('aryan, Knnutcyun Hayastani, 54, whence republished with separate photograph, by Parsegian, ArmenianArchitecture, 7: microfiche 159, C8; additional photographs also published by Parsegian, ArmenianArchitecture, 1: microfiche 36, B7-10 and C 1-2.

dl. 2. 17( 4-Ph9417 I

3.

U VU 1. L (US17htro)8
P.UUU<B> 4./L, UV i ZP'i,

(6I5 1Uq h Ib WLU LAP 1 UlaflQ.Ph4 luIL2 U.U 4rPU, " LP fUI9f UWUlJ Afi6th [lUh]CUl' 8 1IILI ~Ul b]U/ 9.(PhUS)U ObUURAPLhCP .[ U9 PIPCO; 10. IUPuP lP U RS441IU . 11. "' Pb'tIh s.lr.rPUr f U h P4f 9I-P[bf9lPnl7U]
Through the favoring by God of Grigoras monk, this cathedral was built as a refuge for Mazdara (= Mastara). The bride crowned with a cross-shaped crown has as a bridegroom Christ and as the companion of the bridegroom, this Apostle, this Prophet, this martyr, this one keeps Mazdara prosperous and redeems Grigoras 10.4 Mastara: On the basis of a mid-640s date, this equates to 30 November (Fig. 10.4). Located on western section of southern elevation. Copied and published by Aliian, Ayrarat, 136, whence republished, without photograph, by ManHayastani, 54, whence republished, with photograph, by Parsegian, Armenian uE'aryan, K'nnut'yun 7: microfiche 159, C9. Architecture,

4. 5. 6. 7. 8. 9.

86

A CORPUS OF EARLY MEDIEVAL ARMENIAN

INSCRIPTIONS

1. 1 PUln 1UI'Ul7 d, h UPPrI 2. 'b6lR1lMtUSh1B

6ft'U.tb 4. Ph Bh 90USUw4 OqUk!

[-]

3. 8

SUf [.]

[-]

Of the month Arac' [day] 14 at the consecration of this holy church and to the memory of bishop

11. AruL': 24 March 670 (Fig. 11). Panel located on exterior of eastern apse, beneath central window. Copied and published by Mxit'ar Vardapet, "Mi k'ani arjanagrut'iwnner," Handes Hayagitut'iwn 3.2 (1903): 1, whence republished by Kostaneanc', Vimakan Taregir,2; copied and published by Yovsep'ean, K'artiz, no. 9 and pl. 6, whence republished by Strzygowski, Die Baukunst, 1: fig. 40; copied and published by Orbeli, Izbrannyetrudy,402, with separate drawing at 421, whence republished by Abrahamyan, Hayoc'gir ev grr'ut'yun,86, and by Manuc'aryan, K'nnut'yun Hayastani, 100; photographed and published by Stone, Kouymjian, and Lehmann, Album, 114 and Inscription 3.

P U(riPP ulfA.PhOlU 3. ILP4UI L P 4. U2t6 01PI94/PhP ul.UP1hrI'i'ibhP fill PhJ U 5. 6. h 1'4PP, ?flP'11heIlPh P 0P`LP0 7. h1LPUMPUlhllh PhhPP
8.

U 2. Ph IWU-lr f1PLPlI[P]d UPP6hlJ'

1. h 6h f

Uh'PlUSUSIPh I'rLP

41(8{}7n
'U1 RUI

U'iJ4jto, ,

In the 29th year of Constans in the month Mareri which was day 15 this holy cathedral was founded by means of Grigor Mamikonean prince of Armenia and Heline the spouse of the same for the intercession of the builders this 12. T'alin: before 695 (Fig. 12). Four-line inscription, located on a single row of blocks, at roof level. Copied and published by Sahxat'uneanc', Storagrut wn, 2:50; copied and published by Alisan, 2; photographed and published Ayrarat, 138, whence republished by Kostaneanc', VimakanTaregir, no. 5 and and 1; drawn, by Yovsep'ean, Kartiz, pl. photographed, published by Orbeli, Izbrannye 434 and whence trudy, pl. LVIII, republished by Abrahamyan, Hayoc'gir ev grriutyun, 86, and by Hayastani, 87; separate photograph published by Parsegian, ArmenianArManuE'aryan, K'nnut'yun chitecture,1: microfiche 64, D8.

3. aP'b lh bPV UP PII'itL LUlPuBl' iJhlJ 4. ZPILZ LSUl9 lPPhlfl lUPfrl+

1. +bU bUPUUb ILu 11"IPlS t[SPP4 Ph]Plui69 b UP IralPPihbb-8 SIP Ph P 2. 'ba UIP bU 4U1PU GL'Ilh P[UPP[II]98 Pr' U(USflhPLU) h bhLPUMRlPhUllh
PU'/ Uh

I Nerseh apohipatpa[trik] lord of [Si]rak and Alarunik' built this church in the name of the Holy Mother of God for the intercession of me and SuSan my wife and Hrahat our son 13. Naxiavan/Noramanuk: Between ca. 725 and ca. 760 (Fig. 13). Carved onto a tombstone, once located in an underground crypt, now destroyed. Copied and published by Atiian, Sirak, 136, whence republished by Kostaneanc', VimakanTaregir, 3; photographed and published by Yovsep'ean, K'artez,no. 10 and pl. 5; copied, drawn, and pub-

TIMOTHY

GREENWOOD

87

lished, with enhanced photograph, by Orbeli, Izbrannye trudy,411 and pl. LvI.1, whence republished ev and Hayastani, 104. by Manuc'aryan, K'nnut'yun by Abrahamyan, Hayoc'gir gre'ut'yun,86-87,

1. P(USfluI) P{WLb} ruI ZPLWtZ ] Ph (PhUSi)Ufl29 PuLd-u b.bh[[f b1 2. !UYWP UU'I4 hhP PB L'ibJ/h4/IJllta1b ZLLPU 3. UPLL'ZlJh S(1)P LPS1LPLL.9h 4LF<U>9Pt11<>'b11A1LUPPhhIZS i 6 h hgCU14' 4. 4QL9SP1h flJrlPh ZPILLSu .[UL.]B 5. 'USPP CPPtUI UB P UUI PtbUI9 S(UI)flj, Ub
11'IL I'PmtIhP hPUIL',U With an exit commanded by God and a departure called by Christ the blessed Lord Artavazd Kamsarakan apahiwpatpatrik and prince of Armenia, son of Hrahat patrik lord of Sirak and A9arunik', and husband of Suvan Mamikonean was buried in this with his fathers 14. T'alin/Uxtaytur: 15 May 783/14 May 784 (Fig. 14). Located on two blocks forming part of the southwestern pier, interior. Ayrarat,139, whence republished by Kostaneanc', VimakanTaregir, Copied and published by AMihan, 2; photographed and published by Yovs4p'ean, K'artiz, no. 11 and pl. 7, whence republished by Abrahamyan, Hayoc'girev grc'ut'yun,87; separate photograph published by Parsegian, ArmenianArchitecture,1: microfiche 70, C2; photographed and published by Stone, Kouymjian, and Lehmann, Album, 115 and Inscription 4. 6.

,FIPUli

1. 2. 3. 4. 5. 6.

Ukf! 4l/ -, ZU/, Zl[B] k/, 0U ,1UtiLU 1n. 4[11]'tI IlPS.U1SPlP bU ShSP bPuIL9PhU" Ph[P] P ILP4IL'[.] U[B.]ILZfkhhPP1'6hU f [.] "IP IlSU'i h AIikPh'b 8JPUft i W [.11./I b bUW UI'_15-U [t/rIP] Tf-_ 7. bh. bh

SU0~.U 232 of the Armenian era, I Uxtaytur monk and Tuti my brother we brought a water source from K'arkap[.] ... to T'alin. If anyone from ? destroys, may he be liable for my sins

Inscriptionsin Greek
15. Mastara: 480 or undated. Carved onto a single block located on exterior wall. Copied and published by Strzygowski, Die Baukunst, 1:31 and fig. 33, whence republished in Hayastani, 56-58; republished with new suggested reading by E V. ShelovManuL'aryan,K'nnutcyun Kovediaev, "Zametki po grecheskoi epigrafike Armenii," IFZh (1986) 1:66-68. HEPOZ [.]ONTAK aK'or nepo I [p] nep6[ng] .ovt(a) 21 of king Peroz or Peroz Leontak I

[Ae]ovwtK[ng]

16. Ereroyk': Undated (Fig. 16). Carved onto a single block located on eastern side of southern fagade. Copied and published by Strzygowski, Die Baukunst, 1:31 and fig. 32, whence republished by A. arminiennedu IVeau VIesidcle(Paris, 1971), 45; photographed and pubKhatchatrian, L'architecture lished by Shelov-Kovediaev, "Zametki,"59-65.

88 1. 2. 3. 4. 5.

A CORPUS OF EARLY MEDIEVAL ARMENIAN + TQ OIKQ COT HPEHIAFIAC MA KE EIC MAK POTHTAH MEP{E}jQN+ + O)M gaK(5pt)e otZKo ayiIao uo I ph-7x<e>t

INSCRIPTIONS

Holiness befits your house, O Lord, for evermore (Ps. 92 [93]: 5)

U Ia Icp f {e}I0v+ t;SK p6LrILa

17. Ej'miacin: Undated (Fig. 17) Carved onto single block on northern fagade; (a) and (b) flank a circular medallion; beneath the medallion, a tabula ansata engraved with (c); (d) inscribed around circumference of medallion, with additional invocation split between the bars of a cross, itself cut in relief; two doves perched on the medallion, facing each other. arminienne, 67, whence republished, with new Copied and published by Khatchatrian, L'architecture suggested reading, by Shelov-Kovediaev, "Zametki," 68-69; separate photograph published by 1: microfiche 8, B 10. Parsegian, ArmenianArchitecture, (a) (b) (c) (d) KTPI(E) EAEHCONI TON AOTAON I COT APXIAN K(AI)I KT(PIE)EAE(HCON) EAHIAIN AANIHA I TIPEP FAPIKINIC EN TH EKAH(CIA)I{E}HCOT ZIBIOAIN BOH0I HANTACTOYCEYXOMENOTC

I 60v 66v Ioo 'ApXicav (a)KGipt() io1Tov

(b) I"Egnttv i(a0i) Tipep raptdivt; (c) Aavt'1 I
(d) BoTji0t indVaT'0g " (a) (b) (c) (d)
)OL0e

IK(ptw)

AXc(,oov)

VOho

LV

t)

ZptOatv lKrl(oaq) 'If{e}iooi5

Lord, have mercy upon your servant Arkian And Lord have mercy upon Elpis Daniel Tirer Garikinis Help all those who pray in the church Zibithain ofJesus

18. Zvart'noc':After 644 and before 660 (Fig. 18). Found on eastern side of exterior wall. Photographed and published by Strzygowski, Die Baukunst, 1: fig. 30; photographed and published by S. Kh. Mnatsakanian, Zvartnotsi odnotipiye pamiatniki (Erevan, 1971), 11, no. 1 and pl. I, whence 69-70. Shelov-Kovediaev, "Zametki," republished by NAPCHCEHOIHCEN IMNHMONETCATE

Igviovetoamc* Napog AiTnotioEv Narses made. Remember

Inscriptions in Arabic I am very grateful to Dr. Robert Hoyland for studying these inscriptions and supplying his own readings and translations. These differ substantially from those provided by XaE'atryan. These fragments represent those readings which Hoyland considers secure; those which he considered unclear have been excluded. 19. Zvart'noc': 24 December 770-12 December 771 (Fig. 19). Located on plaster surface of a column, interior.

TIMOTHY

GREENWOOD

89

A. A. Xa('atryan, KorpusArabskikh Armeni VIII-XVI vv. (Erevan, 1987), no. 6 and pl. Iv. nadpiesye

[--]
[-]

[-] [-]

. U A= c lJI L4Z

[--]

[-] [--]

and he wrote in the year four and fifty and one hundred 1. [-] [-] [-] b. of Yazid ernorship Jarad and Bgr ys (?) I b. Mariam (?) [-] I [-]

... in the gov-

2. Ilyas b. 'Isa al-Waddaki wrote I [-] May Allah pardon I him and the entirety of his family and his house [-] I [-] and the village I He wrote in the year [-] I 3. In the name ofAllah the beneficent the merciful

I Muhammad b. Abdullth

b. Hamad

20. AruL: 3 July 786-21 June 787. Located on northern elevation, exterior. XaC'atryan,Korpus,no. 4 and pl. II.

[--]
al-Husayn b. Muhammad b. I 'Isa al-Qalsani wrote [in] the year seventy and one hundred

I [-]

B. ARMENIAN MOSAIC INSCRIPTIONS FROM JERUSALEM

1. Musrara/Bird Mosaic: Sixth century? (Fig. 20). Located 350 m north of the Damascus Gate; found in a rectangular room with an apse at the east end. Photographed and published by E J. Bliss and A. C. Dickie, Excavations at Jerusalem 1894-1897 (London, 1898), 253-59; republished, with full bibliography, by M. Avi-Yonah, "Mosaic Pavements in Palestine," QDAP 2 (1932): 171-72 and no. 132; separate photographs published by Der Nersessian, L'artarminien, figs. 44 and 45; photographed and published, with full bibliography, by M. E. Stone, "AReassessment of the Bird and Eustathius Mosaics," in M. E. Stone, R. R. Ervine, and N. Stone, eds., The Armeniansin Jerusalemand the Holy Land, Hebrew University Armenian Studies 4 (Leuven, 2002), 203-19 and fig. 9. 2. UGU' 9PL88 , i UP elP U3Uh) 8(1U)P 91 4hstl

90

A CORPUS OF EARLY MEDIEVAL ARMENIAN

INSCRIPTIONS

For the memory and salvation of all Armenians whose names the Lord knows 2. Mount of Olives/Susannan: Sixth century? (Fig. 21). Located on a pavement in the hall of the Russian Museum on the Mount of Olives. de topographie, Photographed and published by H. Vincent and E M. Abel, Jerusalem. Recherches et vols. and 2 2: XLIII; d'histoire, (Paris, 1926), d'archeologie photographed published by E. Loupl. du mont "Le Musee du russe des Oliviers BIE 13 (1931): 100-101 couvent kianoff, i Jerusalem," and pl. vII; republished by Avi-Yonah, "Mosaic Pavements," 168 and no. 118.

1Usfk 17PUl,

fl[8] flihlZLUGl,

~ WLrhP 4%fh d-e 1UPSlh',LUB

This is the body of Susannan, mother of Artawan Hori 18 3. Mount of Olives/Yakob: Sixth century? (Fig. 22). Located in a chapel north of the Russian church on the Mount of Olives. Photographed and published by Loukianoff, "Le Mus6e," 100 and pl. v.1, whence republished, with full bibliography, by Avi-Yonah, "Mosaic Pavements," 167 and no. 117; separate photograph published by Stone, "Bird and Eustathius Mosaics," fig. 6.

1. UBU .GLUSULL 2. PlrU' S(tLf) BUl84/P0LUB1

4.

h h26* 3. 1P P 6-146"-P(.18

lut

This is the memorial of Lord Yakob which came into being by means of his request 4. Mount of Olives/Valan: Sixth century? (Fig. 23). Located in a vaulted underground tomb in a chapel north of the Russian church on the Mount of Olives. Photographed and published by Yovsip'ean, K'artiz, no. 1 and pl. 1; photographed and published by Loukianoff, "Le Musee," 100 and pl. v.2 (revealing subsequent damage); republished, with full bibliography, by Avi-Yonah, "Mosaic Pavements," 168 and no. 119.

1.

2. U11/U(UShIIU))trRUIhPP 6

fibhllhblfr PLPUILPlU

5. P1118 PhUU fPlfIVh 8 6. flPU1i, U)j1L8 1 .l

7. idUSUULPUZLU ,UU Through having as intercessor before God the holy Isaiah and these blessed fathers I Valan made for the sake of remission of sins this memorial 5. Mount of Olives/T'ewah: Fifth or sixth century. Located between the gate and the Russian house on the Mount of Olives. Researches in Palestineduring the Years 1873C. Clermont-Ganneau, trans. A. Stewart, Archaeological 1874, 2 vols. (London, 1899) 1:329 and pl., whence republished by Avi-Yonah, "Mosaic Pavements," 169 and no. 120.

TIMOTHY

GREENWOOD

91

1. ,4.11- tluL,14'lh#6-hOPl 2. 4llP '-bO A016 I 1UhUlL IL U U6h 3. PUlL WhflahPf

For the prayers and salvation of T'ewah, Abasov, and Murkan 6. Musrara/Ewstat'/Eustathius: After 638 (Fig. 24). Located north of Damascus Gate, in side room of a church, within 100 m of Bird Mosaic (B.1). D. Amit and S. Wolff, "Excavations on an Early Armenian Monastery in the Morasha Neighborhood 26.1-2 (1993): 54; G. Avni, 'Jerusalem as Textbook," BiblicalArchaeological ofJerusalem,' Qadmoniot Review 22.3 (May-June 1996): 40, 42-43; D. Amit and S. Wolff, "AnArmenian Monastery in the Morasha Neighborhood ofJerusalem," in AncientJerusalem Revealed,ed. H. Geva (Jerusalem, 1994), from "The New Armenian M. E. 293-98; Stone, Inscriptions Jerusalem," in ArmenianPerspectives, ed. N. Awde (Richmond, 1997), 263-68; photograph republished in Stone, "Bird and Eustathius Mosaics," fig. 7.

1. 2. 3. 4.

6U bhUSLUPOPY ( P0h hIUHUvIMIXU fP USUltAlf h Slh)lPU bhPU

6. ~I1 U P8h,6I', h U 7. -P(Ph USR)

5. P hWtOPi1 1ff t.JlhP

I Ewstat' elder made this mosaic. Whoever enters this house, remember to Christ me and my brother Lukas